Diffstat (limited to 'noncore/apps/tinykate/libkate/qt3back/qregexp3.cpp') (more/less context) (ignore whitespace changes)
-rw-r--r-- | noncore/apps/tinykate/libkate/qt3back/qregexp3.cpp | 2282 |
1 files changed, 1144 insertions, 1138 deletions
diff --git a/noncore/apps/tinykate/libkate/qt3back/qregexp3.cpp b/noncore/apps/tinykate/libkate/qt3back/qregexp3.cpp index a2c680f..78635b2 100644 --- a/noncore/apps/tinykate/libkate/qt3back/qregexp3.cpp +++ b/noncore/apps/tinykate/libkate/qt3back/qregexp3.cpp | |||
@@ -1,150 +1,156 @@ | |||
1 | /**************************************************************************** | 1 | /**************************************************************************** |
2 | ** $Id$ | 2 | ** $Id$ |
3 | ** | 3 | ** |
4 | ** Implementation of QRegExp class | 4 | ** Implementation of QRegExp class |
5 | ** | 5 | ** |
6 | ** Created : 950126 | 6 | ** Created : 950126 |
7 | ** | 7 | ** |
8 | ** Copyright (C) 1992-2000 Trolltech AS. All rights reserved. | 8 | ** Copyright (C) 1992-2000 Trolltech AS. All rights reserved. |
9 | ** | 9 | ** |
10 | ** This file is part of the tools module of the Qt GUI Toolkit. | 10 | ** This file is part of the tools module of the Qt GUI Toolkit. |
11 | ** | 11 | ** |
12 | ** This file may be distributed under the terms of the Q Public License | 12 | ** This file may be distributed under the terms of the Q Public License |
13 | ** as defined by Trolltech AS of Norway and appearing in the file | 13 | ** as defined by Trolltech AS of Norway and appearing in the file |
14 | ** LICENSE.QPL included in the packaging of this file. | 14 | ** LICENSE.QPL included in the packaging of this file. |
15 | ** | 15 | ** |
16 | ** This file may be distributed and/or modified under the terms of the | 16 | ** This file may be distributed and/or modified under the terms of the |
17 | ** GNU General Public License version 2 as published by the Free Software | 17 | ** GNU General Public License version 2 as published by the Free Software |
18 | ** Foundation and appearing in the file LICENSE.GPL included in the | 18 | ** Foundation and appearing in the file LICENSE.GPL included in the |
19 | ** packaging of this file. | 19 | ** packaging of this file. |
20 | ** | 20 | ** |
21 | ** Licensees holding valid Qt Enterprise Edition or Qt Professional Edition | 21 | ** Licensees holding valid Qt Enterprise Edition or Qt Professional Edition |
22 | ** licenses may use this file in accordance with the Qt Commercial License | 22 | ** licenses may use this file in accordance with the Qt Commercial License |
23 | ** Agreement provided with the Software. | 23 | ** Agreement provided with the Software. |
24 | ** | 24 | ** |
25 | ** This file is provided AS IS with NO WARRANTY OF ANY KIND, INCLUDING THE | 25 | ** This file is provided AS IS with NO WARRANTY OF ANY KIND, INCLUDING THE |
26 | ** WARRANTY OF DESIGN, MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. | 26 | ** WARRANTY OF DESIGN, MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. |
27 | ** | 27 | ** |
28 | ** See http://www.trolltech.com/pricing.html or email sales@trolltech.com for | 28 | ** See http://www.trolltech.com/pricing.html or email sales@trolltech.com for |
29 | ** information about Qt Commercial License Agreements. | 29 | ** information about Qt Commercial License Agreements. |
30 | ** See http://www.trolltech.com/qpl/ for QPL licensing information. | 30 | ** See http://www.trolltech.com/qpl/ for QPL licensing information. |
31 | ** See http://www.trolltech.com/gpl/ for GPL licensing information. | 31 | ** See http://www.trolltech.com/gpl/ for GPL licensing information. |
32 | ** | 32 | ** |
33 | ** Contact info@trolltech.com if any conditions of this licensing are | 33 | ** Contact info@trolltech.com if any conditions of this licensing are |
34 | ** not clear to you. | 34 | ** not clear to you. |
35 | ** | 35 | ** |
36 | **********************************************************************/ | 36 | **********************************************************************/ |
37 | #if QT_VERSION >=300 | 37 | #if QT_VERSION >=300 |
38 | #error QRegExp3 is now in QT 3 use QRegExp instead | 38 | #error QRegExp3 is now in QT 3 use QRegExp instead |
39 | #endif | 39 | #endif |
40 | 40 | ||
41 | #include "qarray.h" | ||
42 | #include "qbitarray.h" | ||
43 | #include "qcache.h" | ||
44 | #include "qintdict.h" | ||
45 | #include "qmap.h" | ||
46 | #if QT_VERSION < 300 | 41 | #if QT_VERSION < 300 |
47 | #include "./qregexp3.h" | 42 | #include "./qregexp3.h" |
48 | #else | 43 | #else |
49 | #include "qregexp.h" | 44 | #include "qregexp.h" |
50 | #endif | 45 | #endif |
51 | #include "qstring.h" | ||
52 | #include "qtl.h" | ||
53 | #include "qvector.h" | ||
54 | 46 | ||
47 | /* OPIE */ | ||
48 | #include <opie2/odebug.h> | ||
49 | |||
50 | /* QT */ | ||
51 | #include <qarray.h> | ||
52 | #include <qbitarray.h> | ||
53 | #include <qcache.h> | ||
54 | #include <qintdict.h> | ||
55 | #include <qmap.h> | ||
56 | #include <qstring.h> | ||
57 | #include <qtl.h> | ||
58 | #include <qvector.h> | ||
59 | |||
60 | /* STD */ | ||
55 | #include <limits.h> | 61 | #include <limits.h> |
56 | 62 | ||
57 | /* | 63 | /* |
58 | WARNING! Be sure to read qregexp.tex before modifying this file. | 64 | WARNING! Be sure to read qregexp.tex before modifying this file. |
59 | */ | 65 | */ |
60 | 66 | ||
61 | /*! | 67 | /*! |
62 | \class QRegExp3 qregexp.h | 68 | \class QRegExp3 qregexp.h |
63 | 69 | ||
64 | \brief The QRegExp class provides pattern matching using regular expressions. | 70 | \brief The QRegExp class provides pattern matching using regular expressions. |
65 | 71 | ||
66 | \ingroup tools | 72 | \ingroup tools |
67 | \ingroup misc | 73 | \ingroup misc |
68 | \ingroup shared | 74 | \ingroup shared |
69 | 75 | ||
70 | 76 | ||
71 | Regular expressions, "regexps", provide a way to find patterns | 77 | Regular expressions, "regexps", provide a way to find patterns |
72 | within text. This is useful in many contexts, for example: | 78 | within text. This is useful in many contexts, for example: |
73 | 79 | ||
74 | <ol> | 80 | <ol> |
75 | <li>\e Validation. A regexp can be used to check whether a piece of | 81 | <li>\e Validation. A regexp can be used to check whether a piece of |
76 | text meets some criteria, e.g. is an integer or contains no | 82 | text meets some criteria, e.g. is an integer or contains no |
77 | whitespace. | 83 | whitespace. |
78 | <li>\e Searching. Regexps provide a much more powerful means of | 84 | <li>\e Searching. Regexps provide a much more powerful means of |
79 | searching text than simple string matching does. For example we can | 85 | searching text than simple string matching does. For example we can |
80 | create a regexp which says "find one of the words 'mail', 'letter' | 86 | create a regexp which says "find one of the words 'mail', 'letter' |
81 | or 'correspondence' but not any of the words 'email', 'mailman' | 87 | or 'correspondence' but not any of the words 'email', 'mailman' |
82 | 'mailer', 'letterbox' etc." | 88 | 'mailer', 'letterbox' etc." |
83 | <li><em>Search and Replace.</em> A regexp can be used to replace a | 89 | <li><em>Search and Replace.</em> A regexp can be used to replace a |
84 | pattern with a piece of text, for example replace all occurrences of | 90 | pattern with a piece of text, for example replace all occurrences of |
85 | '&' with '\&' except where the '&' is already followed by | 91 | '&' with '\&' except where the '&' is already followed by |
86 | 'amp;'. | 92 | 'amp;'. |
87 | <li><em>String Splitting.</em> A regexp can be used to identify | 93 | <li><em>String Splitting.</em> A regexp can be used to identify |
88 | where a string should be split into its component fields, e.g. | 94 | where a string should be split into its component fields, e.g. |
89 | splitting tab delimited strings. | 95 | splitting tab delimited strings. |
90 | </ol> | 96 | </ol> |
91 | 97 | ||
92 | We present a very brief introduction to regexps, a description of | 98 | We present a very brief introduction to regexps, a description of |
93 | Qt's regexp language, some code examples, and finally the function | 99 | Qt's regexp language, some code examples, and finally the function |
94 | documentation. QRegExp is modelled on Perl's regexp engine and fully | 100 | documentation. QRegExp is modelled on Perl's regexp engine and fully |
95 | supports Unicode. QRegExp may also be used in the weaker 'wildcard' | 101 | supports Unicode. QRegExp may also be used in the weaker 'wildcard' |
96 | (globbing) mode which works in a similar way to command shells. A | 102 | (globbing) mode which works in a similar way to command shells. A |
97 | good text on regexps is <i>Mastering Regular Expressions: Powerful | 103 | good text on regexps is <i>Mastering Regular Expressions: Powerful |
98 | Techniques for Perl and Other Tools</i> by Jeffrey E. Friedl, ISBN | 104 | Techniques for Perl and Other Tools</i> by Jeffrey E. Friedl, ISBN |
99 | 1565922573. | 105 | 1565922573. |
100 | 106 | ||
101 | Experienced regexp users may prefer to skip the introduction and | 107 | Experienced regexp users may prefer to skip the introduction and |
102 | go directly to the relevant information: | 108 | go directly to the relevant information: |
103 | 109 | ||
104 | <ul> | 110 | <ul> |
105 | <li><a href="#characters-and-abbreviations-for-sets-of-characters"> | 111 | <li><a href="#characters-and-abbreviations-for-sets-of-characters"> |
106 | Characters and Abbreviations for Sets of Characters</a> | 112 | Characters and Abbreviations for Sets of Characters</a> |
107 | <li><a href="#sets-of-characters">Sets of Characters</a> | 113 | <li><a href="#sets-of-characters">Sets of Characters</a> |
108 | <li><a href="#quantifiers">Quantifiers</a> | 114 | <li><a href="#quantifiers">Quantifiers</a> |
109 | <li><a href="#capturing-text">Capturing Text</a> | 115 | <li><a href="#capturing-text">Capturing Text</a> |
110 | <li><a href="#assertions">Assertions</a> | 116 | <li><a href="#assertions">Assertions</a> |
111 | <li><a href="#wildcard-matching">Wildcard Matching (globbing)</a> | 117 | <li><a href="#wildcard-matching">Wildcard Matching (globbing)</a> |
112 | <li><a href="#perl-users">Notes for Perl Users</a> | 118 | <li><a href="#perl-users">Notes for Perl Users</a> |
113 | <li><a href="#code-examples">Code Examples</a> | 119 | <li><a href="#code-examples">Code Examples</a> |
114 | <li><a href="#member-function-documentation">Member Function Documentation</a> | 120 | <li><a href="#member-function-documentation">Member Function Documentation</a> |
115 | </ul> | 121 | </ul> |
116 | 122 | ||
117 | <b>Introduction</b> | 123 | <b>Introduction</b> |
118 | 124 | ||
119 | Regexps are built up from expressions, quantifiers and assertions. | 125 | Regexps are built up from expressions, quantifiers and assertions. |
120 | The simplest form of expression is simply a character, e.g. <b>x</b> | 126 | The simplest form of expression is simply a character, e.g. <b>x</b> |
121 | or <b>5</b>. An expression can also be a set of characters. For | 127 | or <b>5</b>. An expression can also be a set of characters. For |
122 | example, <b>[ABCD]</b>, will match an <b>A</b> or a <b>B</b> or a | 128 | example, <b>[ABCD]</b>, will match an <b>A</b> or a <b>B</b> or a |
123 | <b>C</b> or a <b>D</b>. As a shorthand we could write this as | 129 | <b>C</b> or a <b>D</b>. As a shorthand we could write this as |
124 | <b>[A-D]</b>. If we want to match any of the captital letters in the | 130 | <b>[A-D]</b>. If we want to match any of the captital letters in the |
125 | English alphabet we can write <b>[A-Z]</b>. A quantifier tells the | 131 | English alphabet we can write <b>[A-Z]</b>. A quantifier tells the |
126 | regexp engine how many occurrences of the expression we want, e.g. | 132 | regexp engine how many occurrences of the expression we want, e.g. |
127 | <b>x{1,1}</b> means match an <b>x</b> which occurs at least once and | 133 | <b>x{1,1}</b> means match an <b>x</b> which occurs at least once and |
128 | at most once. We'll look at assertions and more complex expressions | 134 | at most once. We'll look at assertions and more complex expressions |
129 | later. Note that regexps cannot be used to check for balanced | 135 | later. Note that regexps cannot be used to check for balanced |
130 | brackets or tags (unless you know the maximum level of nesting). | 136 | brackets or tags (unless you know the maximum level of nesting). |
131 | 137 | ||
132 | 138 | ||
133 | We'll start by writing a regexp to match integers in the range 0 to | 139 | We'll start by writing a regexp to match integers in the range 0 to |
134 | 99. We will require at least one digit so we will start with | 140 | 99. We will require at least one digit so we will start with |
135 | <b>[0-9]{1,1}</b> which means match a digit exactly once. This | 141 | <b>[0-9]{1,1}</b> which means match a digit exactly once. This |
136 | regexp alone will match integers in the range 0 to 9. To match one | 142 | regexp alone will match integers in the range 0 to 9. To match one |
137 | or two digits we can increase the maximum number of occurrences so | 143 | or two digits we can increase the maximum number of occurrences so |
138 | the regexp becomes <b>[0-9]{1,2}</b> meaning match a digit at least | 144 | the regexp becomes <b>[0-9]{1,2}</b> meaning match a digit at least |
139 | once and at most twice. However, this regexp as it stands will not | 145 | once and at most twice. However, this regexp as it stands will not |
140 | match correctly. This regexp will match one or two digits \e within | 146 | match correctly. This regexp will match one or two digits \e within |
141 | a string. To ensure that we match against the whole string we must | 147 | a string. To ensure that we match against the whole string we must |
142 | use the anchor assertions. We need <b>^</b> (caret) which when it is | 148 | use the anchor assertions. We need <b>^</b> (caret) which when it is |
143 | the first character in the regexp means that the regexp must match | 149 | the first character in the regexp means that the regexp must match |
144 | from the beginning of the string. And we also need <b>$</b> (dollar) | 150 | from the beginning of the string. And we also need <b>$</b> (dollar) |
145 | which when it is the last character in the regexp means that the | 151 | which when it is the last character in the regexp means that the |
146 | regexp must match until the end of the string. So now our regexp is | 152 | regexp must match until the end of the string. So now our regexp is |
147 | <b>^[0-9]{1,2}$</b>. Note that assertions do not match any | 153 | <b>^[0-9]{1,2}$</b>. Note that assertions do not match any |
148 | characters. | 154 | characters. |
149 | 155 | ||
150 | If you've seen regexps elsewhere they may have looked different from | 156 | If you've seen regexps elsewhere they may have looked different from |
@@ -422,3346 +428,3346 @@ | |||
422 | '#define' followed by at least one whitespace followed by at least | 428 | '#define' followed by at least one whitespace followed by at least |
423 | one word character optionally followed by whitespace. This regexp | 429 | one word character optionally followed by whitespace. This regexp |
424 | will match define's that exist but have no value, i.e. it will not | 430 | will match define's that exist but have no value, i.e. it will not |
425 | match '#define INTMAX 32767' but it will match '#define <u>DEBUG</u>'. | 431 | match '#define INTMAX 32767' but it will match '#define <u>DEBUG</u>'. |
426 | 432 | ||
427 | </ul> | 433 | </ul> |
428 | 434 | ||
429 | <a name="wildcard-matching"><b>Wildcard Matching (globbing)</b></a> | 435 | <a name="wildcard-matching"><b>Wildcard Matching (globbing)</b></a> |
430 | 436 | ||
431 | Most command shells such as \e bash or \e cmd support "file | 437 | Most command shells such as \e bash or \e cmd support "file |
432 | globbing", the ability to identify a group of files by using | 438 | globbing", the ability to identify a group of files by using |
433 | wildcards. Wildcard matching is much simpler than full regexps and | 439 | wildcards. Wildcard matching is much simpler than full regexps and |
434 | has only four features: | 440 | has only four features: |
435 | 441 | ||
436 | <ul> | 442 | <ul> |
437 | 443 | ||
438 | <li><b>c</b> Any character represents itself apart from those | 444 | <li><b>c</b> Any character represents itself apart from those |
439 | mentioned below. Thus <b>c</b> matches the character \e c. | 445 | mentioned below. Thus <b>c</b> matches the character \e c. |
440 | 446 | ||
441 | <li><b>?</b> This matches any single character. It is the same as | 447 | <li><b>?</b> This matches any single character. It is the same as |
442 | <b>.</b> in full regexps. | 448 | <b>.</b> in full regexps. |
443 | 449 | ||
444 | <li><b>*</b> This matches zero or more of any characters. It is the | 450 | <li><b>*</b> This matches zero or more of any characters. It is the |
445 | same as <b>.*</b> in full regexps. | 451 | same as <b>.*</b> in full regexps. |
446 | 452 | ||
447 | <li><b>[...]</b> Sets of characters can be represented in square | 453 | <li><b>[...]</b> Sets of characters can be represented in square |
448 | brackets the same as for full regexps. | 454 | brackets the same as for full regexps. |
449 | 455 | ||
450 | <!-- JASMIN: Are the character classes, \w, etc supported in | 456 | <!-- JASMIN: Are the character classes, \w, etc supported in |
451 | wildcards? --> | 457 | wildcards? --> |
452 | 458 | ||
453 | </ul> | 459 | </ul> |
454 | 460 | ||
455 | For example if we are in wildcard mode and have strings which | 461 | For example if we are in wildcard mode and have strings which |
456 | contain filenames we could identify HTML files with <b>*.html</b>. | 462 | contain filenames we could identify HTML files with <b>*.html</b>. |
457 | This will match zero or more characters followed by a dot followed | 463 | This will match zero or more characters followed by a dot followed |
458 | by 'h', 't', 'm' and 'l'. | 464 | by 'h', 't', 'm' and 'l'. |
459 | 465 | ||
460 | <a name="perl-users"><b>Notes for Perl Users</b></a> | 466 | <a name="perl-users"><b>Notes for Perl Users</b></a> |
461 | 467 | ||
462 | Most of the character class abbreviations supported by Perl are | 468 | Most of the character class abbreviations supported by Perl are |
463 | supported by QRegExp, see | 469 | supported by QRegExp, see |
464 | <a href="#characters-and-abbreviations-for-sets-of-characters"> | 470 | <a href="#characters-and-abbreviations-for-sets-of-characters"> |
465 | characters and abbreviations for sets of characters</a>. | 471 | characters and abbreviations for sets of characters</a>. |
466 | 472 | ||
467 | QRegExp's quantifiers are the same as Perl's greedy quantifiers. | 473 | QRegExp's quantifiers are the same as Perl's greedy quantifiers. |
468 | Non-greedy matching cannot be applied to individual quantifiers, but | 474 | Non-greedy matching cannot be applied to individual quantifiers, but |
469 | can be applied to all the quantifiers in the pattern. For example, | 475 | can be applied to all the quantifiers in the pattern. For example, |
470 | to match the Perl regex <b>ro+?m</b> requires: | 476 | to match the Perl regex <b>ro+?m</b> requires: |
471 | \code | 477 | \code |
472 | QRegExp rx( "ro+m" ); | 478 | QRegExp rx( "ro+m" ); |
473 | rx.setMinimal( TRUE ); | 479 | rx.setMinimal( TRUE ); |
474 | \endcode | 480 | \endcode |
475 | 481 | ||
476 | The equivalent of Perl's <tt>/i</tt> option is | 482 | The equivalent of Perl's <tt>/i</tt> option is |
477 | setCaseSensitive(FALSE). | 483 | setCaseSensitive(FALSE). |
478 | 484 | ||
479 | Perl's <tt>/g</tt> option can be emulated using a | 485 | Perl's <tt>/g</tt> option can be emulated using a |
480 | <a href="#cap_in_a_loop">loop</a>. | 486 | <a href="#cap_in_a_loop">loop</a>. |
481 | 487 | ||
482 | In QRegExp <b>.</b> matches any character, therefore all QRegExp | 488 | In QRegExp <b>.</b> matches any character, therefore all QRegExp |
483 | regexps have the equivalent of Perl's <tt>/s</tt> option. QRegExp | 489 | regexps have the equivalent of Perl's <tt>/s</tt> option. QRegExp |
484 | does not have an equivalent to Perl's <tt>/m</tt> option, but this | 490 | does not have an equivalent to Perl's <tt>/m</tt> option, but this |
485 | can be emulated in various ways for example by splitting the input | 491 | can be emulated in various ways for example by splitting the input |
486 | into lines or by looping with a regexp that searches for newlines. | 492 | into lines or by looping with a regexp that searches for newlines. |
487 | 493 | ||
488 | Because QRegExp is string oriented there are no \A, \Z or \z | 494 | Because QRegExp is string oriented there are no \A, \Z or \z |
489 | assertions. The \G assertion is not supported but can be emulated in | 495 | assertions. The \G assertion is not supported but can be emulated in |
490 | a loop. | 496 | a loop. |
491 | 497 | ||
492 | Perl's $& is cap(0) or capturedTexts()[0]. There are no QRegExp | 498 | Perl's $& is cap(0) or capturedTexts()[0]. There are no QRegExp |
493 | equivalents for $`, $' or $+. $1, $2 etc correspond to | 499 | equivalents for $`, $' or $+. $1, $2 etc correspond to |
494 | cap(1) or capturedTexts()[1], cap(2) or capturedTexts()[2], etc. | 500 | cap(1) or capturedTexts()[1], cap(2) or capturedTexts()[2], etc. |
495 | 501 | ||
496 | To substitute a pattern use QString::replace(). | 502 | To substitute a pattern use QString::replace(). |
497 | 503 | ||
498 | Perl's extended <tt>/x</tt> syntax is not supported, nor are regexp | 504 | Perl's extended <tt>/x</tt> syntax is not supported, nor are regexp |
499 | comments (?#comment) or directives, e.g. (?i). | 505 | comments (?#comment) or directives, e.g. (?i). |
500 | 506 | ||
501 | Both zero-width positive and zero-width negative lookahead | 507 | Both zero-width positive and zero-width negative lookahead |
502 | assertions (?=pattern) and (?!pattern) are supported with the same | 508 | assertions (?=pattern) and (?!pattern) are supported with the same |
503 | syntax as Perl. Perl's lookbehind assertions, "independent" | 509 | syntax as Perl. Perl's lookbehind assertions, "independent" |
504 | subexpressions and conditional expressions are not supported. | 510 | subexpressions and conditional expressions are not supported. |
505 | 511 | ||
506 | Non-capturing parenthesis are also supported, with the same | 512 | Non-capturing parenthesis are also supported, with the same |
507 | (?:pattern) syntax. | 513 | (?:pattern) syntax. |
508 | 514 | ||
509 | See QStringList::split() and QStringList::join() for equivalents to | 515 | See QStringList::split() and QStringList::join() for equivalents to |
510 | Perl's split and join functions. | 516 | Perl's split and join functions. |
511 | 517 | ||
512 | Note: because C++ transforms \\'s they must be written \e twice in | 518 | Note: because C++ transforms \\'s they must be written \e twice in |
513 | code, e.g. <b>\\</b><b>b</b> must be written <b>\\</b><b>\\</b><b>b</b>. | 519 | code, e.g. <b>\\</b><b>b</b> must be written <b>\\</b><b>\\</b><b>b</b>. |
514 | 520 | ||
515 | <a name="code-examples"><b>Code Examples</b></a> | 521 | <a name="code-examples"><b>Code Examples</b></a> |
516 | 522 | ||
517 | \code | 523 | \code |
518 | QRegExp rx( "^\\d\\d?$" );// Match integers 0 to 99 | 524 | QRegExp rx( "^\\d\\d?$" ); // Match integers 0 to 99 |
519 | rx.search( "123" ); // Returns -1 (no match) | 525 | rx.search( "123" ); // Returns -1 (no match) |
520 | rx.search( "-6" ); // Returns -1 (no match) | 526 | rx.search( "-6" ); // Returns -1 (no match) |
521 | rx.search( "6" ); // Returns 0 (matched as position 0) | 527 | rx.search( "6" ); // Returns 0 (matched as position 0) |
522 | \endcode | 528 | \endcode |
523 | 529 | ||
524 | The third string matches '<u>6</u>'. This is a simple validation | 530 | The third string matches '<u>6</u>'. This is a simple validation |
525 | regexp for integers in the range 0 to 99. | 531 | regexp for integers in the range 0 to 99. |
526 | 532 | ||
527 | \code | 533 | \code |
528 | QRegExp rx( "^\\S+$" );// Match strings which have no whitespace | 534 | QRegExp rx( "^\\S+$" ); // Match strings which have no whitespace |
529 | rx.search( "Hello world" );// Returns -1 (no match) | 535 | rx.search( "Hello world" ); // Returns -1 (no match) |
530 | rx.search( "This_is-OK" );// Returns 0 (matched at position 0) | 536 | rx.search( "This_is-OK" ); // Returns 0 (matched at position 0) |
531 | \endcode | 537 | \endcode |
532 | 538 | ||
533 | The second string matches '<u>This_is-OK</u>'. We've used the | 539 | The second string matches '<u>This_is-OK</u>'. We've used the |
534 | character set abbreviation '\S' (non-whitespace) and the anchors to | 540 | character set abbreviation '\S' (non-whitespace) and the anchors to |
535 | match strings which contain no whitespace. | 541 | match strings which contain no whitespace. |
536 | 542 | ||
537 | In the following example we match strings containing 'mail' or | 543 | In the following example we match strings containing 'mail' or |
538 | 'letter' or 'correspondence' but only match whole words i.e. not | 544 | 'letter' or 'correspondence' but only match whole words i.e. not |
539 | 'email' | 545 | 'email' |
540 | 546 | ||
541 | \code | 547 | \code |
542 | QRegExp rx( "\\b(mail|letter|correspondence)\\b" ); | 548 | QRegExp rx( "\\b(mail|letter|correspondence)\\b" ); |
543 | rx.search( "I sent you an email" ); // Returns -1 (no match) | 549 | rx.search( "I sent you an email" ); // Returns -1 (no match) |
544 | rx.search( "Please write the letter" ); // Returns 17 (matched at position 17) | 550 | rx.search( "Please write the letter" ); // Returns 17 (matched at position 17) |
545 | \endcode | 551 | \endcode |
546 | 552 | ||
547 | The second string matches "Please write the <u>letter</u>". The word | 553 | The second string matches "Please write the <u>letter</u>". The word |
548 | 'letter' is also captured (because of the parenthesis). We can see | 554 | 'letter' is also captured (because of the parenthesis). We can see |
549 | what text we've captured like this: | 555 | what text we've captured like this: |
550 | 556 | ||
551 | \code | 557 | \code |
552 | QString captured = rx.cap( 1 ); // captured contains "letter" | 558 | QString captured = rx.cap( 1 ); // captured contains "letter" |
553 | \endcode | 559 | \endcode |
554 | 560 | ||
555 | This will capture the text from the first set of capturing | 561 | This will capture the text from the first set of capturing |
556 | parenthesis (counting capturing left parenthesis from left to | 562 | parenthesis (counting capturing left parenthesis from left to |
557 | right). The parenthesis are counted from 1 since cap( 0 ) is the | 563 | right). The parenthesis are counted from 1 since cap( 0 ) is the |
558 | whole matched regexp (equivalent to '&' in most regexp engines). | 564 | whole matched regexp (equivalent to '&' in most regexp engines). |
559 | 565 | ||
560 | \code | 566 | \code |
561 | QRegExp rx( "&(?!amp;)" ); // Match ampersands but not & | 567 | QRegExp rx( "&(?!amp;)" ); // Match ampersands but not & |
562 | QString line1 = "This & that"; | 568 | QString line1 = "This & that"; |
563 | line1.replace( rx, "&" ); | 569 | line1.replace( rx, "&" ); |
564 | // line1 == "This & that" | 570 | // line1 == "This & that" |
565 | QString line2 = "His & hers & theirs"; | 571 | QString line2 = "His & hers & theirs"; |
566 | line2.replace( rx, "&" ); | 572 | line2.replace( rx, "&" ); |
567 | // line2 == "His & hers & theirs" | 573 | // line2 == "His & hers & theirs" |
568 | \endcode | 574 | \endcode |
569 | 575 | ||
570 | Here we've passed the QRegExp to QString's replace() function to | 576 | Here we've passed the QRegExp to QString's replace() function to |
571 | replace the matched text with new text. | 577 | replace the matched text with new text. |
572 | 578 | ||
573 | \code | 579 | \code |
574 | QString str = "One Eric another Eirik, and an Ericsson. How many Eiriks, Eric?"; | 580 | QString str = "One Eric another Eirik, and an Ericsson. How many Eiriks, Eric?"; |
575 | QRegExp rx( "\\b(Eric|Eirik)\\b" );// Match Eric or Eirik | 581 | QRegExp rx( "\\b(Eric|Eirik)\\b" ); // Match Eric or Eirik |
576 | int pos = 0; // Where we are in the string | 582 | int pos = 0; // Where we are in the string |
577 | int count = 0; // How many Eric and Eirik's we've counted | 583 | int count = 0; // How many Eric and Eirik's we've counted |
578 | while ( pos >= 0 ) { | 584 | while ( pos >= 0 ) { |
579 | pos = rx.search( str, pos ); | 585 | pos = rx.search( str, pos ); |
580 | if ( pos >= 0 ) { | 586 | if ( pos >= 0 ) { |
581 | pos++;// Move along in str | 587 | pos++; // Move along in str |
582 | count++;// Count our Eric or Eirik | 588 | count++; // Count our Eric or Eirik |
583 | } | 589 | } |
584 | } | 590 | } |
585 | \endcode | 591 | \endcode |
586 | 592 | ||
587 | We've used the search() function to repeatedly match the regexp in | 593 | We've used the search() function to repeatedly match the regexp in |
588 | the string. Note that instead of moving forward by one character at | 594 | the string. Note that instead of moving forward by one character at |
589 | a time <tt>pos++</tt> we could have written <tt>pos += | 595 | a time <tt>pos++</tt> we could have written <tt>pos += |
590 | rx.matchedLength()</tt> to skip over the already matched string. The | 596 | rx.matchedLength()</tt> to skip over the already matched string. The |
591 | count will equal 3, matching 'One <u>Eric</u> another <u>Eirik</u>, | 597 | count will equal 3, matching 'One <u>Eric</u> another <u>Eirik</u>, |
592 | and an Ericsson. How many Eiriks, <u>Eric</u>?'; it doesn't match | 598 | and an Ericsson. How many Eiriks, <u>Eric</u>?'; it doesn't match |
593 | 'Ericsson' or 'Eiriks' because they are not bounded by non-word | 599 | 'Ericsson' or 'Eiriks' because they are not bounded by non-word |
594 | boundaries. | 600 | boundaries. |
595 | 601 | ||
596 | One common use of regexps is to split lines of delimited data into | 602 | One common use of regexps is to split lines of delimited data into |
597 | their component fields. | 603 | their component fields. |
598 | 604 | ||
599 | \code | 605 | \code |
600 | str = "Trolltech AS\twww.trolltech.com\tNorway"; | 606 | str = "Trolltech AS\twww.trolltech.com\tNorway"; |
601 | QString company, web, country; | 607 | QString company, web, country; |
602 | rx.setPattern( "^([^\t]+)\t([^\t]+)\t([^\t]+)$" ); | 608 | rx.setPattern( "^([^\t]+)\t([^\t]+)\t([^\t]+)$" ); |
603 | if ( rx.search( str ) != -1 ) { | 609 | if ( rx.search( str ) != -1 ) { |
604 | company = rx.cap( 1 ); | 610 | company = rx.cap( 1 ); |
605 | web= rx.cap( 2 ); | 611 | web = rx.cap( 2 ); |
606 | country = rx.cap( 3 ); | 612 | country = rx.cap( 3 ); |
607 | } | 613 | } |
608 | \endcode | 614 | \endcode |
609 | 615 | ||
610 | In this example our input lines have the format company name, web | 616 | In this example our input lines have the format company name, web |
611 | address and country. Unfortunately the regexp is rather long and not | 617 | address and country. Unfortunately the regexp is rather long and not |
612 | very versatile -- the code will break if we add any more fields. A | 618 | very versatile -- the code will break if we add any more fields. A |
613 | simpler and better solution is to look for the separator, '\t' in | 619 | simpler and better solution is to look for the separator, '\t' in |
614 | this case, and take the surrounding text. The QStringList split() | 620 | this case, and take the surrounding text. The QStringList split() |
615 | function can take a separator string or regexp as an argument and | 621 | function can take a separator string or regexp as an argument and |
616 | split a string accordingly. | 622 | split a string accordingly. |
617 | 623 | ||
618 | \code | 624 | \code |
619 | QStringList field = QStringList::split( "\t", str ); | 625 | QStringList field = QStringList::split( "\t", str ); |
620 | \endcode | 626 | \endcode |
621 | 627 | ||
622 | Here field[0] is the company, field[1] the web address and so on. | 628 | Here field[0] is the company, field[1] the web address and so on. |
623 | 629 | ||
624 | To immitate the matching of a shell we can use wildcard mode. | 630 | To immitate the matching of a shell we can use wildcard mode. |
625 | 631 | ||
626 | \code | 632 | \code |
627 | QRegExp rx( "*.html" );// Invalid regexp: * doesn't quantify anything | 633 | QRegExp rx( "*.html" ); // Invalid regexp: * doesn't quantify anything |
628 | rx.setWildcard( TRUE );// Now its a valid wildcard regexp | 634 | rx.setWildcard( TRUE ); // Now its a valid wildcard regexp |
629 | rx.search( "index.html" );// Returns 0 (matched at position 0) | 635 | rx.search( "index.html" ); // Returns 0 (matched at position 0) |
630 | rx.search( "default.htm" );// Returns -1 (no match) | 636 | rx.search( "default.htm" ); // Returns -1 (no match) |
631 | rx.search( "readme.txt" );// Returns -1 (no match) | 637 | rx.search( "readme.txt" ); // Returns -1 (no match) |
632 | \endcode | 638 | \endcode |
633 | 639 | ||
634 | Wildcard matching can be convenient because of its simplicity, but | 640 | Wildcard matching can be convenient because of its simplicity, but |
635 | any wildcard regex can be defined using full regexps, e.g. | 641 | any wildcard regex can be defined using full regexps, e.g. |
636 | <b>.*\.html$</b>. Notice that we can't match both \c .html and \c | 642 | <b>.*\.html$</b>. Notice that we can't match both \c .html and \c |
637 | .htm files with a wildcard unless we use <b>*.htm*</b> which will | 643 | .htm files with a wildcard unless we use <b>*.htm*</b> which will |
638 | also match 'test.html.bak'. A full regexp gives us the precision we | 644 | also match 'test.html.bak'. A full regexp gives us the precision we |
639 | need, <b>.*\.html?$</b>. | 645 | need, <b>.*\.html?$</b>. |
640 | 646 | ||
641 | QRegExp can match case insensitively using setCaseSensitive(), and | 647 | QRegExp can match case insensitively using setCaseSensitive(), and |
642 | can use non-greedy matching, see setMinimal(). By default QRegExp | 648 | can use non-greedy matching, see setMinimal(). By default QRegExp |
643 | uses full regexps but this can be changed with setWildcard(). | 649 | uses full regexps but this can be changed with setWildcard(). |
644 | Searching can be forward with search() or backward with searchRev(). | 650 | Searching can be forward with search() or backward with searchRev(). |
645 | Captured text can be accessed using capturedTexts() which returns a | 651 | Captured text can be accessed using capturedTexts() which returns a |
646 | string list of all captured strings, or using cap() which returns | 652 | string list of all captured strings, or using cap() which returns |
647 | the captured string for the given index. The pos() function takes a | 653 | the captured string for the given index. The pos() function takes a |
648 | match index and returns the position in the string where the match | 654 | match index and returns the position in the string where the match |
649 | was made (or -1 if there was no match). | 655 | was made (or -1 if there was no match). |
650 | 656 | ||
651 | \sa QRegExpValidator QString QStringList | 657 | \sa QRegExpValidator QString QStringList |
652 | 658 | ||
653 | <a name="member-function-documentation"/> | 659 | <a name="member-function-documentation"/> |
654 | */ | 660 | */ |
655 | 661 | ||
656 | static const int NumBadChars = 128; | 662 | static const int NumBadChars = 128; |
657 | #define BadChar( ch ) ( (ch).cell() % NumBadChars ) | 663 | #define BadChar( ch ) ( (ch).cell() % NumBadChars ) |
658 | 664 | ||
659 | static const int NoOccurrence = INT_MAX; | 665 | static const int NoOccurrence = INT_MAX; |
660 | static const int EmptyCapture = INT_MAX; | 666 | static const int EmptyCapture = INT_MAX; |
661 | static const int InftyLen = INT_MAX; | 667 | static const int InftyLen = INT_MAX; |
662 | static const int InftyRep = 1000; | 668 | static const int InftyRep = 1000; |
663 | static const int EOS = -1; | 669 | static const int EOS = -1; |
664 | 670 | ||
665 | #ifndef QT_NO_REGEXP_OPTIM | 671 | #ifndef QT_NO_REGEXP_OPTIM |
666 | static int engCount = 0; | 672 | static int engCount = 0; |
667 | static QArray<int> *noOccurrences = 0; | 673 | static QArray<int> *noOccurrences = 0; |
668 | static QArray<int> *firstOccurrenceAtZero = 0; | 674 | static QArray<int> *firstOccurrenceAtZero = 0; |
669 | #endif | 675 | #endif |
670 | 676 | ||
671 | /* | 677 | /* |
672 | Merges two QArrays of ints and puts the result into the first one. | 678 | Merges two QArrays of ints and puts the result into the first one. |
673 | */ | 679 | */ |
674 | static void mergeInto( QArray<int> *a, const QArray<int>& b ) | 680 | static void mergeInto( QArray<int> *a, const QArray<int>& b ) |
675 | { | 681 | { |
676 | int asize = a->size(); | 682 | int asize = a->size(); |
677 | int bsize = b.size(); | 683 | int bsize = b.size(); |
678 | if ( asize == 0 ) { | 684 | if ( asize == 0 ) { |
679 | *a = b.copy(); | 685 | *a = b.copy(); |
680 | #ifndef QT_NO_REGEXP_OPTIM | 686 | #ifndef QT_NO_REGEXP_OPTIM |
681 | } else if ( bsize == 1 && (*a)[asize - 1] < b[0] ) { | 687 | } else if ( bsize == 1 && (*a)[asize - 1] < b[0] ) { |
682 | a->resize( asize + 1 ); | 688 | a->resize( asize + 1 ); |
683 | (*a)[asize] = b[0]; | 689 | (*a)[asize] = b[0]; |
684 | #endif | 690 | #endif |
685 | } else if ( bsize >= 1 ) { | 691 | } else if ( bsize >= 1 ) { |
686 | int csize = asize + bsize; | 692 | int csize = asize + bsize; |
687 | QArray<int> c( csize ); | 693 | QArray<int> c( csize ); |
688 | int i = 0, j = 0, k = 0; | 694 | int i = 0, j = 0, k = 0; |
689 | while ( i < asize ) { | 695 | while ( i < asize ) { |
690 | if ( j < bsize ) { | 696 | if ( j < bsize ) { |
691 | if ( (*a)[i] == b[j] ) { | 697 | if ( (*a)[i] == b[j] ) { |
692 | i++; | 698 | i++; |
693 | csize--; | 699 | csize--; |
694 | } else if ( (*a)[i] < b[j] ) { | 700 | } else if ( (*a)[i] < b[j] ) { |
695 | c[k++] = (*a)[i++]; | 701 | c[k++] = (*a)[i++]; |
696 | } else { | 702 | } else { |
697 | c[k++] = b[j++]; | 703 | c[k++] = b[j++]; |
698 | } | 704 | } |
699 | } else { | 705 | } else { |
700 | memcpy( c.data() + k, (*a).data() + i, | 706 | memcpy( c.data() + k, (*a).data() + i, |
701 | (asize - i) * sizeof(int) ); | 707 | (asize - i) * sizeof(int) ); |
702 | break; | 708 | break; |
703 | } | 709 | } |
704 | } | 710 | } |
705 | c.resize( csize ); | 711 | c.resize( csize ); |
706 | if ( j < bsize ) | 712 | if ( j < bsize ) |
707 | memcpy( c.data() + k, b.data() + j, (bsize - j) * sizeof(int) ); | 713 | memcpy( c.data() + k, b.data() + j, (bsize - j) * sizeof(int) ); |
708 | *a = c; | 714 | *a = c; |
709 | } | 715 | } |
710 | } | 716 | } |
711 | 717 | ||
712 | /* | 718 | /* |
713 | Merges two disjoint QMaps of (int, int) pairs and puts the result into the | 719 | Merges two disjoint QMaps of (int, int) pairs and puts the result into the |
714 | first one. | 720 | first one. |
715 | */ | 721 | */ |
716 | static void mergeInto( QMap<int, int> *a, const QMap<int, int>& b ) | 722 | static void mergeInto( QMap<int, int> *a, const QMap<int, int>& b ) |
717 | { | 723 | { |
718 | QMap<int, int>::ConstIterator it; | 724 | QMap<int, int>::ConstIterator it; |
719 | for ( it = b.begin(); it != b.end(); ++it ) | 725 | for ( it = b.begin(); it != b.end(); ++it ) |
720 | a->insert( it.key(), *it ); | 726 | a->insert( it.key(), *it ); |
721 | } | 727 | } |
722 | 728 | ||
723 | /* | 729 | /* |
724 | Returns the value associated to key k in QMap m of (int, int) pairs, or 0 if | 730 | Returns the value associated to key k in QMap m of (int, int) pairs, or 0 if |
725 | no such value is explicitly present. | 731 | no such value is explicitly present. |
726 | */ | 732 | */ |
727 | static int at( const QMap<int, int>& m, int k ) | 733 | static int at( const QMap<int, int>& m, int k ) |
728 | { | 734 | { |
729 | QMap<int, int>::ConstIterator it = m.find( k ); | 735 | QMap<int, int>::ConstIterator it = m.find( k ); |
730 | if ( it == m.end() ) | 736 | if ( it == m.end() ) |
731 | return 0; | 737 | return 0; |
732 | else | 738 | else |
733 | return *it; | 739 | return *it; |
734 | } | 740 | } |
735 | 741 | ||
736 | #ifndef QT_NO_REGEXP_WILDCARD | 742 | #ifndef QT_NO_REGEXP_WILDCARD |
737 | /* | 743 | /* |
738 | Translates a wildcard pattern to an equivalent regular expression pattern | 744 | Translates a wildcard pattern to an equivalent regular expression pattern |
739 | (e.g., *.cpp to .*\.cpp). | 745 | (e.g., *.cpp to .*\.cpp). |
740 | */ | 746 | */ |
741 | static QString wc2rx( const QString& wc ) | 747 | static QString wc2rx( const QString& wc ) |
742 | { | 748 | { |
743 | int wclen = wc.length(); | 749 | int wclen = wc.length(); |
744 | QString rx = QString::fromLatin1( "" ); | 750 | QString rx = QString::fromLatin1( "" ); |
745 | int i = 0; | 751 | int i = 0; |
746 | while ( i < wclen ) { | 752 | while ( i < wclen ) { |
747 | QChar c = wc[i++]; | 753 | QChar c = wc[i++]; |
748 | switch ( c.unicode() ) { | 754 | switch ( c.unicode() ) { |
749 | case '*': | 755 | case '*': |
750 | rx += QString::fromLatin1( ".*" ); | 756 | rx += QString::fromLatin1( ".*" ); |
751 | break; | 757 | break; |
752 | case '?': | 758 | case '?': |
753 | rx += QChar( '.' ); | 759 | rx += QChar( '.' ); |
754 | break; | 760 | break; |
755 | case '$': | 761 | case '$': |
756 | case '(': | 762 | case '(': |
757 | case ')': | 763 | case ')': |
758 | case '+': | 764 | case '+': |
759 | case '.': | 765 | case '.': |
760 | case '\\': | 766 | case '\\': |
761 | case '^': | 767 | case '^': |
762 | case '{': | 768 | case '{': |
763 | case '|': | 769 | case '|': |
764 | case '}': | 770 | case '}': |
765 | rx += QChar( '\\' ); | 771 | rx += QChar( '\\' ); |
766 | rx += c; | 772 | rx += c; |
767 | break; | 773 | break; |
768 | case '[': | 774 | case '[': |
769 | rx += c; | 775 | rx += c; |
770 | if ( wc[i] == QChar('^') ) | 776 | if ( wc[i] == QChar('^') ) |
771 | rx += wc[i++]; | 777 | rx += wc[i++]; |
772 | if ( i < wclen ) { | 778 | if ( i < wclen ) { |
773 | if ( rx[i] == ']' ) | 779 | if ( rx[i] == ']' ) |
774 | rx += wc[i++]; | 780 | rx += wc[i++]; |
775 | while ( i < wclen && wc[i] != QChar(']') ) { | 781 | while ( i < wclen && wc[i] != QChar(']') ) { |
776 | if ( wc[i] == '\\' ) | 782 | if ( wc[i] == '\\' ) |
777 | rx += QChar( '\\' ); | 783 | rx += QChar( '\\' ); |
778 | rx += wc[i++]; | 784 | rx += wc[i++]; |
779 | } | 785 | } |
780 | } | 786 | } |
781 | break; | 787 | break; |
782 | default: | 788 | default: |
783 | rx += c; | 789 | rx += c; |
784 | } | 790 | } |
785 | } | 791 | } |
786 | return rx; | 792 | return rx; |
787 | } | 793 | } |
788 | #endif | 794 | #endif |
789 | 795 | ||
790 | /* | 796 | /* |
791 | The class QRegExpEngine encapsulates a modified nondeterministic finite | 797 | The class QRegExpEngine encapsulates a modified nondeterministic finite |
792 | automaton (NFA). | 798 | automaton (NFA). |
793 | */ | 799 | */ |
794 | class QRegExpEngine : public QShared | 800 | class QRegExpEngine : public QShared |
795 | { | 801 | { |
796 | public: | 802 | public: |
797 | #ifndef QT_NO_REGEXP_CCLASS | 803 | #ifndef QT_NO_REGEXP_CCLASS |
798 | /* | 804 | /* |
799 | The class CharClass represents a set of characters, such as can be found | 805 | The class CharClass represents a set of characters, such as can be found |
800 | in regular expressions (e.g., [a-z] denotes the set {a, b, ..., z}). | 806 | in regular expressions (e.g., [a-z] denotes the set {a, b, ..., z}). |
801 | */ | 807 | */ |
802 | class CharClass | 808 | class CharClass |
803 | { | 809 | { |
804 | public: | 810 | public: |
805 | CharClass(); | 811 | CharClass(); |
806 | CharClass( const CharClass& cc ) { operator=( cc ); } | 812 | CharClass( const CharClass& cc ) { operator=( cc ); } |
807 | 813 | ||
808 | CharClass& operator=( const CharClass& cc ); | 814 | CharClass& operator=( const CharClass& cc ); |
809 | 815 | ||
810 | void clear(); | 816 | void clear(); |
811 | bool negative() const { return n; } | 817 | bool negative() const { return n; } |
812 | void setNegative( bool negative ); | 818 | void setNegative( bool negative ); |
813 | void addCategories( int cats ); | 819 | void addCategories( int cats ); |
814 | void addRange( ushort from, ushort to ); | 820 | void addRange( ushort from, ushort to ); |
815 | void addSingleton( ushort ch ) { addRange( ch, ch ); } | 821 | void addSingleton( ushort ch ) { addRange( ch, ch ); } |
816 | 822 | ||
817 | bool in( QChar ch ) const; | 823 | bool in( QChar ch ) const; |
818 | #ifndef QT_NO_REGEXP_OPTIM | 824 | #ifndef QT_NO_REGEXP_OPTIM |
819 | const QArray<int>& firstOccurrence() const { return occ1; } | 825 | const QArray<int>& firstOccurrence() const { return occ1; } |
820 | #endif | 826 | #endif |
821 | 827 | ||
822 | #if defined(QT_DEBUG) | 828 | #if defined(QT_DEBUG) |
823 | void dump() const; | 829 | void dump() const; |
824 | #endif | 830 | #endif |
825 | 831 | ||
826 | private: | 832 | private: |
827 | /* | 833 | /* |
828 | The struct Range represents a range of characters (e.g., [0-9] denotes | 834 | The struct Range represents a range of characters (e.g., [0-9] denotes |
829 | range 48 to 57). | 835 | range 48 to 57). |
830 | */ | 836 | */ |
831 | struct Range | 837 | struct Range |
832 | { | 838 | { |
833 | ushort from; // 48 | 839 | ushort from; // 48 |
834 | ushort to; // 57 | 840 | ushort to; // 57 |
835 | }; | 841 | }; |
836 | 842 | ||
837 | int c; // character classes | 843 | int c; // character classes |
838 | QArray<Range> r; // character ranges | 844 | QArray<Range> r; // character ranges |
839 | bool n; // negative? | 845 | bool n; // negative? |
840 | #ifndef QT_NO_REGEXP_OPTIM | 846 | #ifndef QT_NO_REGEXP_OPTIM |
841 | QArray<int> occ1; // first-occurrence array | 847 | QArray<int> occ1; // first-occurrence array |
842 | #endif | 848 | #endif |
843 | }; | 849 | }; |
844 | #else | 850 | #else |
845 | struct CharClass | 851 | struct CharClass |
846 | { | 852 | { |
847 | int x; // dummy | 853 | int x; // dummy |
848 | 854 | ||
849 | #ifndef QT_NO_REGEXP_OPTIM | 855 | #ifndef QT_NO_REGEXP_OPTIM |
850 | const QArray<int>& firstOccurrence() const { | 856 | const QArray<int>& firstOccurrence() const { |
851 | return *firstOccurrenceAtZero; | 857 | return *firstOccurrenceAtZero; |
852 | } | 858 | } |
853 | #endif | 859 | #endif |
854 | }; | 860 | }; |
855 | #endif | 861 | #endif |
856 | 862 | ||
857 | QRegExpEngine( bool caseSensitive ) { setup( caseSensitive ); } | 863 | QRegExpEngine( bool caseSensitive ) { setup( caseSensitive ); } |
858 | QRegExpEngine( const QString& rx, bool caseSensitive ); | 864 | QRegExpEngine( const QString& rx, bool caseSensitive ); |
859 | #ifndef QT_NO_REGEXP_OPTIM | 865 | #ifndef QT_NO_REGEXP_OPTIM |
860 | ~QRegExpEngine(); | 866 | ~QRegExpEngine(); |
861 | #endif | 867 | #endif |
862 | 868 | ||
863 | bool isValid() const { return valid; } | 869 | bool isValid() const { return valid; } |
864 | bool caseSensitive() const { return cs; } | 870 | bool caseSensitive() const { return cs; } |
865 | int numCaptures() const { return realncap; } | 871 | int numCaptures() const { return realncap; } |
866 | QArray<int> match( const QString& str, int pos, bool minimal, | 872 | QArray<int> match( const QString& str, int pos, bool minimal, |
867 | bool oneTest ); | 873 | bool oneTest ); |
868 | int matchedLength() const { return mmMatchedLen; } | 874 | int matchedLength() const { return mmMatchedLen; } |
869 | 875 | ||
870 | int createState( QChar ch ); | 876 | int createState( QChar ch ); |
871 | int createState( const CharClass& cc ); | 877 | int createState( const CharClass& cc ); |
872 | #ifndef QT_NO_REGEXP_BACKREF | 878 | #ifndef QT_NO_REGEXP_BACKREF |
873 | int createState( int bref ); | 879 | int createState( int bref ); |
874 | #endif | 880 | #endif |
875 | 881 | ||
876 | void addCatTransitions( const QArray<int>& from, const QArray<int>& to ); | 882 | void addCatTransitions( const QArray<int>& from, const QArray<int>& to ); |
877 | #ifndef QT_NO_REGEXP_CAPTURE | 883 | #ifndef QT_NO_REGEXP_CAPTURE |
878 | void addPlusTransitions( const QArray<int>& from, const QArray<int>& to, | 884 | void addPlusTransitions( const QArray<int>& from, const QArray<int>& to, |
879 | int atom ); | 885 | int atom ); |
880 | #endif | 886 | #endif |
881 | 887 | ||
882 | #ifndef QT_NO_REGEXP_ANCHOR_ALT | 888 | #ifndef QT_NO_REGEXP_ANCHOR_ALT |
883 | int anchorAlternation( int a, int b ); | 889 | int anchorAlternation( int a, int b ); |
884 | int anchorConcatenation( int a, int b ); | 890 | int anchorConcatenation( int a, int b ); |
885 | #else | 891 | #else |
886 | int anchorAlternation( int a, int b ) { return a & b; } | 892 | int anchorAlternation( int a, int b ) { return a & b; } |
887 | int anchorConcatenation( int a, int b ) { return a | b; } | 893 | int anchorConcatenation( int a, int b ) { return a | b; } |
888 | #endif | 894 | #endif |
889 | void addAnchors( int from, int to, int a ); | 895 | void addAnchors( int from, int to, int a ); |
890 | 896 | ||
891 | #ifndef QT_NO_REGEXP_OPTIM | 897 | #ifndef QT_NO_REGEXP_OPTIM |
892 | void setupGoodStringHeuristic( int earlyStart, int lateStart, | 898 | void setupGoodStringHeuristic( int earlyStart, int lateStart, |
893 | const QString& str ); | 899 | const QString& str ); |
894 | void setupBadCharHeuristic( int minLen, const QArray<int>& firstOcc ); | 900 | void setupBadCharHeuristic( int minLen, const QArray<int>& firstOcc ); |
895 | void heuristicallyChooseHeuristic(); | 901 | void heuristicallyChooseHeuristic(); |
896 | #endif | 902 | #endif |
897 | 903 | ||
898 | #if defined(QT_DEBUG) | 904 | #if defined(QT_DEBUG) |
899 | void dump() const; | 905 | void dump() const; |
900 | #endif | 906 | #endif |
901 | 907 | ||
902 | private: | 908 | private: |
903 | enum { CharClassBit = 0x10000, BackRefBit = 0x20000 }; | 909 | enum { CharClassBit = 0x10000, BackRefBit = 0x20000 }; |
904 | 910 | ||
905 | /* | 911 | /* |
906 | The struct State represents one state in a modified NFA. The input | 912 | The struct State represents one state in a modified NFA. The input |
907 | characters matched are stored in the state instead of on the transitions, | 913 | characters matched are stored in the state instead of on the transitions, |
908 | something possible for an automaton constructed from a regular expression. | 914 | something possible for an automaton constructed from a regular expression. |
909 | */ | 915 | */ |
910 | struct State | 916 | struct State |
911 | { | 917 | { |
912 | #ifndef QT_NO_REGEXP_CAPTURE | 918 | #ifndef QT_NO_REGEXP_CAPTURE |
913 | int atom; // which atom does this state belong to? | 919 | int atom; // which atom does this state belong to? |
914 | #endif | 920 | #endif |
915 | int match; // what does it match? (see CharClassBit and BackRefBit) | 921 | int match; // what does it match? (see CharClassBit and BackRefBit) |
916 | QArray<int> outs; // out-transitions | 922 | QArray<int> outs; // out-transitions |
917 | QMap<int, int> *reenter; // atoms reentered when transiting out | 923 | QMap<int, int> *reenter; // atoms reentered when transiting out |
918 | QMap<int, int> *anchors; // anchors met when transiting out | 924 | QMap<int, int> *anchors; // anchors met when transiting out |
919 | 925 | ||
920 | #ifndef QT_NO_REGEXP_CAPTURE | 926 | #ifndef QT_NO_REGEXP_CAPTURE |
921 | State( int a, int m ) | 927 | State( int a, int m ) |
922 | : atom( a ), match( m ), reenter( 0 ), anchors( 0 ) { } | 928 | : atom( a ), match( m ), reenter( 0 ), anchors( 0 ) { } |
923 | #else | 929 | #else |
924 | State( int m ) | 930 | State( int m ) |
925 | : match( m ), reenter( 0 ), anchors( 0 ) { } | 931 | : match( m ), reenter( 0 ), anchors( 0 ) { } |
926 | #endif | 932 | #endif |
927 | ~State() { delete reenter; delete anchors; } | 933 | ~State() { delete reenter; delete anchors; } |
928 | }; | 934 | }; |
929 | 935 | ||
930 | #ifndef QT_NO_REGEXP_LOOKAHEAD | 936 | #ifndef QT_NO_REGEXP_LOOKAHEAD |
931 | /* | 937 | /* |
932 | The struct Lookahead represents a lookahead a la Perl (e.g., (?=foo) and | 938 | The struct Lookahead represents a lookahead a la Perl (e.g., (?=foo) and |
933 | (?!bar)). | 939 | (?!bar)). |
934 | */ | 940 | */ |
935 | struct Lookahead | 941 | struct Lookahead |
936 | { | 942 | { |
937 | QRegExpEngine *eng; // NFA representing the embedded regular expression | 943 | QRegExpEngine *eng; // NFA representing the embedded regular expression |
938 | bool neg; // negative lookahead? | 944 | bool neg; // negative lookahead? |
939 | 945 | ||
940 | Lookahead( QRegExpEngine *eng0, bool neg0 ) | 946 | Lookahead( QRegExpEngine *eng0, bool neg0 ) |
941 | : eng( eng0 ), neg( neg0 ) { } | 947 | : eng( eng0 ), neg( neg0 ) { } |
942 | ~Lookahead() { delete eng; } | 948 | ~Lookahead() { delete eng; } |
943 | }; | 949 | }; |
944 | #endif | 950 | #endif |
945 | 951 | ||
946 | #ifndef QT_NO_REGEXP_CAPTURE | 952 | #ifndef QT_NO_REGEXP_CAPTURE |
947 | /* | 953 | /* |
948 | The struct Atom represents one node in the hierarchy of regular expression | 954 | The struct Atom represents one node in the hierarchy of regular expression |
949 | atoms. | 955 | atoms. |
950 | */ | 956 | */ |
951 | struct Atom | 957 | struct Atom |
952 | { | 958 | { |
953 | int parent; // index of parent in array of atoms | 959 | int parent; // index of parent in array of atoms |
954 | int capture; // index of capture, from 1 to ncap | 960 | int capture; // index of capture, from 1 to ncap |
955 | }; | 961 | }; |
956 | #endif | 962 | #endif |
957 | 963 | ||
958 | #ifndef QT_NO_REGEXP_ANCHOR_ALT | 964 | #ifndef QT_NO_REGEXP_ANCHOR_ALT |
959 | /* | 965 | /* |
960 | The struct AnchorAlternation represents a pair of anchors with OR | 966 | The struct AnchorAlternation represents a pair of anchors with OR |
961 | semantics. | 967 | semantics. |
962 | */ | 968 | */ |
963 | struct AnchorAlternation | 969 | struct AnchorAlternation |
964 | { | 970 | { |
965 | int a; // this anchor... | 971 | int a; // this anchor... |
966 | int b; // ...or this one | 972 | int b; // ...or this one |
967 | }; | 973 | }; |
968 | #endif | 974 | #endif |
969 | 975 | ||
970 | enum { InitialState = 0, FinalState = 1 }; | 976 | enum { InitialState = 0, FinalState = 1 }; |
971 | void setup( bool caseSensitive ); | 977 | void setup( bool caseSensitive ); |
972 | int setupState( int match ); | 978 | int setupState( int match ); |
973 | 979 | ||
974 | /* | 980 | /* |
975 | Let's hope that 13 lookaheads and 14 back-references are enough. | 981 | Let's hope that 13 lookaheads and 14 back-references are enough. |
976 | */ | 982 | */ |
977 | enum { MaxLookaheads = 13, MaxBackRefs = 14 }; | 983 | enum { MaxLookaheads = 13, MaxBackRefs = 14 }; |
978 | enum { Anchor_Dollar = 0x00000001, Anchor_Caret = 0x00000002, | 984 | enum { Anchor_Dollar = 0x00000001, Anchor_Caret = 0x00000002, |
979 | Anchor_Word = 0x00000004, Anchor_NonWord = 0x00000008, | 985 | Anchor_Word = 0x00000004, Anchor_NonWord = 0x00000008, |
980 | Anchor_FirstLookahead = 0x00000010, | 986 | Anchor_FirstLookahead = 0x00000010, |
981 | Anchor_BackRef1Empty = Anchor_FirstLookahead << MaxLookaheads, | 987 | Anchor_BackRef1Empty = Anchor_FirstLookahead << MaxLookaheads, |
982 | Anchor_BackRef0Empty = Anchor_BackRef1Empty >> 1, | 988 | Anchor_BackRef0Empty = Anchor_BackRef1Empty >> 1, |
983 | Anchor_Alternation = Anchor_BackRef1Empty << MaxBackRefs, | 989 | Anchor_Alternation = Anchor_BackRef1Empty << MaxBackRefs, |
984 | 990 | ||
985 | Anchor_LookaheadMask = ( Anchor_FirstLookahead - 1 ) ^ | 991 | Anchor_LookaheadMask = ( Anchor_FirstLookahead - 1 ) ^ |
986 | ( (Anchor_FirstLookahead << MaxLookaheads) - 1 ) }; | 992 | ( (Anchor_FirstLookahead << MaxLookaheads) - 1 ) }; |
987 | #ifndef QT_NO_REGEXP_CAPTURE | 993 | #ifndef QT_NO_REGEXP_CAPTURE |
988 | int startAtom( bool capture ); | 994 | int startAtom( bool capture ); |
989 | void finishAtom( int atom ) { cf = f[atom].parent; } | 995 | void finishAtom( int atom ) { cf = f[atom].parent; } |
990 | #endif | 996 | #endif |
991 | 997 | ||
992 | #ifndef QT_NO_REGEXP_LOOKAHEAD | 998 | #ifndef QT_NO_REGEXP_LOOKAHEAD |
993 | int addLookahead( QRegExpEngine *eng, bool negative ); | 999 | int addLookahead( QRegExpEngine *eng, bool negative ); |
994 | #endif | 1000 | #endif |
995 | 1001 | ||
996 | #ifndef QT_NO_REGEXP_CAPTURE | 1002 | #ifndef QT_NO_REGEXP_CAPTURE |
997 | bool isBetterCapture( const int *begin1, const int *end1, const int *begin2, | 1003 | bool isBetterCapture( const int *begin1, const int *end1, const int *begin2, |
998 | const int *end2 ); | 1004 | const int *end2 ); |
999 | #endif | 1005 | #endif |
1000 | bool testAnchor( int i, int a, const int *capBegin ); | 1006 | bool testAnchor( int i, int a, const int *capBegin ); |
1001 | 1007 | ||
1002 | #ifndef QT_NO_REGEXP_OPTIM | 1008 | #ifndef QT_NO_REGEXP_OPTIM |
1003 | bool goodStringMatch(); | 1009 | bool goodStringMatch(); |
1004 | bool badCharMatch(); | 1010 | bool badCharMatch(); |
1005 | #else | 1011 | #else |
1006 | bool bruteMatch(); | 1012 | bool bruteMatch(); |
1007 | #endif | 1013 | #endif |
1008 | bool matchHere(); | 1014 | bool matchHere(); |
1009 | 1015 | ||
1010 | QVector<State> s; // array of states | 1016 | QVector<State> s; // array of states |
1011 | int ns; // number of states | 1017 | int ns; // number of states |
1012 | #ifndef QT_NO_REGEXP_CAPTURE | 1018 | #ifndef QT_NO_REGEXP_CAPTURE |
1013 | QArray<Atom> f; // atom hierarchy | 1019 | QArray<Atom> f; // atom hierarchy |
1014 | int nf; // number of atoms | 1020 | int nf; // number of atoms |
1015 | int cf; // current atom | 1021 | int cf; // current atom |
1016 | #endif | 1022 | #endif |
1017 | int realncap; // number of captures, seen from the outside | 1023 | int realncap; // number of captures, seen from the outside |
1018 | int ncap; // number of captures, seen from the inside | 1024 | int ncap; // number of captures, seen from the inside |
1019 | #ifndef QT_NO_REGEXP_CCLASS | 1025 | #ifndef QT_NO_REGEXP_CCLASS |
1020 | QVector<CharClass> cl; // array of character classes | 1026 | QVector<CharClass> cl; // array of character classes |
1021 | #endif | 1027 | #endif |
1022 | #ifndef QT_NO_REGEXP_LOOKAHEAD | 1028 | #ifndef QT_NO_REGEXP_LOOKAHEAD |
1023 | QVector<Lookahead> ahead; // array of lookaheads | 1029 | QVector<Lookahead> ahead; // array of lookaheads |
1024 | #endif | 1030 | #endif |
1025 | #ifndef QT_NO_REGEXP_ANCHOR_ALT | 1031 | #ifndef QT_NO_REGEXP_ANCHOR_ALT |
1026 | QArray<AnchorAlternation> aa; // array of (a, b) pairs of anchors | 1032 | QArray<AnchorAlternation> aa; // array of (a, b) pairs of anchors |
1027 | #endif | 1033 | #endif |
1028 | #ifndef QT_NO_REGEXP_OPTIM | 1034 | #ifndef QT_NO_REGEXP_OPTIM |
1029 | bool caretAnchored; // does the regexp start with ^? | 1035 | bool caretAnchored; // does the regexp start with ^? |
1030 | #endif | 1036 | #endif |
1031 | bool valid; // is the regular expression valid? | 1037 | bool valid; // is the regular expression valid? |
1032 | bool cs; // case sensitive? | 1038 | bool cs; // case sensitive? |
1033 | #ifndef QT_NO_REGEXP_BACKREF | 1039 | #ifndef QT_NO_REGEXP_BACKREF |
1034 | int nbrefs; // number of back-references | 1040 | int nbrefs; // number of back-references |
1035 | #endif | 1041 | #endif |
1036 | 1042 | ||
1037 | #ifndef QT_NO_REGEXP_OPTIM | 1043 | #ifndef QT_NO_REGEXP_OPTIM |
1038 | bool useGoodStringHeuristic; // use goodStringMatch? otherwise badCharMatch | 1044 | bool useGoodStringHeuristic; // use goodStringMatch? otherwise badCharMatch |
1039 | 1045 | ||
1040 | int goodEarlyStart; // the index where goodStr can first occur in a match | 1046 | int goodEarlyStart; // the index where goodStr can first occur in a match |
1041 | int goodLateStart; // the index where goodStr can last occur in a match | 1047 | int goodLateStart; // the index where goodStr can last occur in a match |
1042 | QString goodStr; // the string that any match has to contain | 1048 | QString goodStr; // the string that any match has to contain |
1043 | 1049 | ||
1044 | int minl; // the minimum length of a match | 1050 | int minl; // the minimum length of a match |
1045 | QArray<int> occ1; // first-occurrence array | 1051 | QArray<int> occ1; // first-occurrence array |
1046 | #endif | 1052 | #endif |
1047 | 1053 | ||
1048 | /* | 1054 | /* |
1049 | The class Box is an abstraction for a regular expression fragment. It can | 1055 | The class Box is an abstraction for a regular expression fragment. It can |
1050 | also be seen as one node in the syntax tree of a regular expression with | 1056 | also be seen as one node in the syntax tree of a regular expression with |
1051 | synthetized attributes. | 1057 | synthetized attributes. |
1052 | 1058 | ||
1053 | It's interface is ugly for performance reasons. | 1059 | It's interface is ugly for performance reasons. |
1054 | */ | 1060 | */ |
1055 | class Box | 1061 | class Box |
1056 | { | 1062 | { |
1057 | public: | 1063 | public: |
1058 | Box( QRegExpEngine *engine ); | 1064 | Box( QRegExpEngine *engine ); |
1059 | Box( const Box& b ) { operator=( b ); } | 1065 | Box( const Box& b ) { operator=( b ); } |
1060 | 1066 | ||
1061 | Box& operator=( const Box& b ); | 1067 | Box& operator=( const Box& b ); |
1062 | 1068 | ||
1063 | void clear() { operator=(Box(eng)); } | 1069 | void clear() { operator=(Box(eng)); } |
1064 | void set( QChar ch ); | 1070 | void set( QChar ch ); |
1065 | void set( const CharClass& cc ); | 1071 | void set( const CharClass& cc ); |
1066 | #ifndef QT_NO_REGEXP_BACKREF | 1072 | #ifndef QT_NO_REGEXP_BACKREF |
1067 | void set( int bref ); | 1073 | void set( int bref ); |
1068 | #endif | 1074 | #endif |
1069 | 1075 | ||
1070 | void cat( const Box& b ); | 1076 | void cat( const Box& b ); |
1071 | void orx( const Box& b ); | 1077 | void orx( const Box& b ); |
1072 | void plus( int atom ); | 1078 | void plus( int atom ); |
1073 | void opt(); | 1079 | void opt(); |
1074 | void catAnchor( int a ); | 1080 | void catAnchor( int a ); |
1075 | #ifndef QT_NO_REGEXP_OPTIM | 1081 | #ifndef QT_NO_REGEXP_OPTIM |
1076 | void setupHeuristics(); | 1082 | void setupHeuristics(); |
1077 | #endif | 1083 | #endif |
1078 | 1084 | ||
1079 | #if defined(QT_DEBUG) | 1085 | #if defined(QT_DEBUG) |
1080 | void dump() const; | 1086 | void dump() const; |
1081 | #endif | 1087 | #endif |
1082 | 1088 | ||
1083 | private: | 1089 | private: |
1084 | void addAnchorsToEngine( const Box& to ) const; | 1090 | void addAnchorsToEngine( const Box& to ) const; |
1085 | 1091 | ||
1086 | QRegExpEngine *eng; // the automaton under construction | 1092 | QRegExpEngine *eng; // the automaton under construction |
1087 | QArray<int> ls; // the left states (firstpos) | 1093 | QArray<int> ls; // the left states (firstpos) |
1088 | QArray<int> rs; // the right states (lastpos) | 1094 | QArray<int> rs; // the right states (lastpos) |
1089 | QMap<int, int> lanchors; // the left anchors | 1095 | QMap<int, int> lanchors; // the left anchors |
1090 | QMap<int, int> ranchors; // the right anchors | 1096 | QMap<int, int> ranchors; // the right anchors |
1091 | int skipanchors; // the anchors to match if the box is skipped | 1097 | int skipanchors; // the anchors to match if the box is skipped |
1092 | 1098 | ||
1093 | #ifndef QT_NO_REGEXP_OPTIM | 1099 | #ifndef QT_NO_REGEXP_OPTIM |
1094 | int earlyStart; // the index where str can first occur | 1100 | int earlyStart; // the index where str can first occur |
1095 | int lateStart; // the index where str can last occur | 1101 | int lateStart; // the index where str can last occur |
1096 | QString str; // a string that has to occur in any match | 1102 | QString str; // a string that has to occur in any match |
1097 | QString leftStr; // a string occurring at the left of this box | 1103 | QString leftStr; // a string occurring at the left of this box |
1098 | QString rightStr; // a string occurring at the right of this box | 1104 | QString rightStr; // a string occurring at the right of this box |
1099 | int maxl; // the maximum length of this box (possibly InftyLen) | 1105 | int maxl; // the maximum length of this box (possibly InftyLen) |
1100 | #endif | 1106 | #endif |
1101 | 1107 | ||
1102 | int minl; // the minimum length of this box | 1108 | int minl; // the minimum length of this box |
1103 | #ifndef QT_NO_REGEXP_OPTIM | 1109 | #ifndef QT_NO_REGEXP_OPTIM |
1104 | QArray<int> occ1; // first-occurrence array | 1110 | QArray<int> occ1; // first-occurrence array |
1105 | #endif | 1111 | #endif |
1106 | }; | 1112 | }; |
1107 | friend class Box; | 1113 | friend class Box; |
1108 | 1114 | ||
1109 | /* | 1115 | /* |
1110 | This is the lexical analyzer for regular expressions. | 1116 | This is the lexical analyzer for regular expressions. |
1111 | */ | 1117 | */ |
1112 | enum { Tok_Eos, Tok_Dollar, Tok_LeftParen, Tok_MagicLeftParen, | 1118 | enum { Tok_Eos, Tok_Dollar, Tok_LeftParen, Tok_MagicLeftParen, |
1113 | Tok_PosLookahead, Tok_NegLookahead, Tok_RightParen, Tok_CharClass, | 1119 | Tok_PosLookahead, Tok_NegLookahead, Tok_RightParen, Tok_CharClass, |
1114 | Tok_Caret, Tok_Quantifier, Tok_Bar, Tok_Word, Tok_NonWord, | 1120 | Tok_Caret, Tok_Quantifier, Tok_Bar, Tok_Word, Tok_NonWord, |
1115 | Tok_Char = 0x10000, Tok_BackRef = 0x20000 }; | 1121 | Tok_Char = 0x10000, Tok_BackRef = 0x20000 }; |
1116 | int getChar(); | 1122 | int getChar(); |
1117 | int getEscape(); | 1123 | int getEscape(); |
1118 | #ifndef QT_NO_REGEXP_INTERVAL | 1124 | #ifndef QT_NO_REGEXP_INTERVAL |
1119 | int getRep( int def ); | 1125 | int getRep( int def ); |
1120 | #endif | 1126 | #endif |
1121 | #ifndef QT_NO_REGEXP_LOOKAHEAD | 1127 | #ifndef QT_NO_REGEXP_LOOKAHEAD |
1122 | void skipChars( int n ); | 1128 | void skipChars( int n ); |
1123 | #endif | 1129 | #endif |
1124 | void startTokenizer( const QChar *rx, int len ); | 1130 | void startTokenizer( const QChar *rx, int len ); |
1125 | int getToken(); | 1131 | int getToken(); |
1126 | 1132 | ||
1127 | const QChar *yyIn; // a pointer to the input regular expression pattern | 1133 | const QChar *yyIn; // a pointer to the input regular expression pattern |
1128 | int yyPos0; // the position of yyTok in the input pattern | 1134 | int yyPos0; // the position of yyTok in the input pattern |
1129 | int yyPos; // the position of the next character to read | 1135 | int yyPos; // the position of the next character to read |
1130 | int yyLen; // the length of yyIn | 1136 | int yyLen; // the length of yyIn |
1131 | int yyCh; // the last character read | 1137 | int yyCh; // the last character read |
1132 | CharClass *yyCharClass; // attribute for Tok_CharClass tokens | 1138 | CharClass *yyCharClass; // attribute for Tok_CharClass tokens |
1133 | int yyMinRep; // attribute for Tok_Quantifier | 1139 | int yyMinRep; // attribute for Tok_Quantifier |
1134 | int yyMaxRep; // ditto | 1140 | int yyMaxRep; // ditto |
1135 | bool yyError; // syntax error or overflow during parsing? | 1141 | bool yyError; // syntax error or overflow during parsing? |
1136 | 1142 | ||
1137 | /* | 1143 | /* |
1138 | This is the syntactic analyzer for regular expressions. | 1144 | This is the syntactic analyzer for regular expressions. |
1139 | */ | 1145 | */ |
1140 | int parse( const QChar *rx, int len ); | 1146 | int parse( const QChar *rx, int len ); |
1141 | void parseAtom( Box *box ); | 1147 | void parseAtom( Box *box ); |
1142 | void parseFactor( Box *box ); | 1148 | void parseFactor( Box *box ); |
1143 | void parseTerm( Box *box ); | 1149 | void parseTerm( Box *box ); |
1144 | void parseExpression( Box *box ); | 1150 | void parseExpression( Box *box ); |
1145 | 1151 | ||
1146 | int yyTok; // the last token read | 1152 | int yyTok; // the last token read |
1147 | bool yyMayCapture; // set this to FALSE to disable capturing | 1153 | bool yyMayCapture; // set this to FALSE to disable capturing |
1148 | 1154 | ||
1149 | /* | 1155 | /* |
1150 | This is the engine state during matching. | 1156 | This is the engine state during matching. |
1151 | */ | 1157 | */ |
1152 | const QString *mmStr; // a pointer to the input QString | 1158 | const QString *mmStr; // a pointer to the input QString |
1153 | const QChar *mmIn; // a pointer to the input string data | 1159 | const QChar *mmIn; // a pointer to the input string data |
1154 | int mmPos; // the current position in the string | 1160 | int mmPos; // the current position in the string |
1155 | int mmLen; // the length of the input string | 1161 | int mmLen; // the length of the input string |
1156 | bool mmMinimal; // minimal matching? | 1162 | bool mmMinimal; // minimal matching? |
1157 | QArray<int> mmCaptured; // an array of pairs (start, len) | 1163 | QArray<int> mmCaptured; // an array of pairs (start, len) |
1158 | QArray<int> mmCapturedNoMatch; // an array of pairs (-1, -1) | 1164 | QArray<int> mmCapturedNoMatch; // an array of pairs (-1, -1) |
1159 | QArray<int> mmBigArray; // big QArray<int> array | 1165 | QArray<int> mmBigArray; // big QArray<int> array |
1160 | int *mmInNextStack; // is state is mmNextStack? | 1166 | int *mmInNextStack; // is state is mmNextStack? |
1161 | int *mmCurStack; // stack of current states | 1167 | int *mmCurStack; // stack of current states |
1162 | int *mmNextStack; // stack of next states | 1168 | int *mmNextStack; // stack of next states |
1163 | int *mmCurCapBegin; // start of current states' captures | 1169 | int *mmCurCapBegin; // start of current states' captures |
1164 | int *mmNextCapBegin; // start of next states' captures | 1170 | int *mmNextCapBegin; // start of next states' captures |
1165 | int *mmCurCapEnd; // end of current states' captures | 1171 | int *mmCurCapEnd; // end of current states' captures |
1166 | int *mmNextCapEnd; // end of next states' captures | 1172 | int *mmNextCapEnd; // end of next states' captures |
1167 | int *mmTempCapBegin; // start of temporary captures | 1173 | int *mmTempCapBegin; // start of temporary captures |
1168 | int *mmTempCapEnd; // end of temporary captures | 1174 | int *mmTempCapEnd; // end of temporary captures |
1169 | int *mmCapBegin; // start of captures for a next state | 1175 | int *mmCapBegin; // start of captures for a next state |
1170 | int *mmCapEnd; // end of captures for a next state | 1176 | int *mmCapEnd; // end of captures for a next state |
1171 | int *mmSlideTab; // bump-along slide table for bad-character heuristic | 1177 | int *mmSlideTab; // bump-along slide table for bad-character heuristic |
1172 | int mmSlideTabSize; // size of slide table | 1178 | int mmSlideTabSize; // size of slide table |
1173 | #ifndef QT_NO_REGEXP_BACKREF | 1179 | #ifndef QT_NO_REGEXP_BACKREF |
1174 | QIntDict<int> mmSleeping; // dictionary of back-reference sleepers | 1180 | QIntDict<int> mmSleeping; // dictionary of back-reference sleepers |
1175 | #endif | 1181 | #endif |
1176 | int mmMatchedLen; // length of match or of matched string for partial match | 1182 | int mmMatchedLen; // length of match or of matched string for partial match |
1177 | }; | 1183 | }; |
1178 | 1184 | ||
1179 | QRegExpEngine::QRegExpEngine( const QString& rx, bool caseSensitive ) | 1185 | QRegExpEngine::QRegExpEngine( const QString& rx, bool caseSensitive ) |
1180 | #ifndef QT_NO_REGEXP_BACKREF | 1186 | #ifndef QT_NO_REGEXP_BACKREF |
1181 | : mmSleeping( 101 ) | 1187 | : mmSleeping( 101 ) |
1182 | #endif | 1188 | #endif |
1183 | { | 1189 | { |
1184 | setup( caseSensitive ); | 1190 | setup( caseSensitive ); |
1185 | valid = ( parse(rx.unicode(), rx.length()) == (int) rx.length() ); | 1191 | valid = ( parse(rx.unicode(), rx.length()) == (int) rx.length() ); |
1186 | } | 1192 | } |
1187 | 1193 | ||
1188 | #ifndef QT_NO_REGEXP_OPTIM | 1194 | #ifndef QT_NO_REGEXP_OPTIM |
1189 | QRegExpEngine::~QRegExpEngine() | 1195 | QRegExpEngine::~QRegExpEngine() |
1190 | { | 1196 | { |
1191 | if ( --engCount == 0 ) { | 1197 | if ( --engCount == 0 ) { |
1192 | delete noOccurrences; | 1198 | delete noOccurrences; |
1193 | noOccurrences = 0; | 1199 | noOccurrences = 0; |
1194 | delete firstOccurrenceAtZero; | 1200 | delete firstOccurrenceAtZero; |
1195 | firstOccurrenceAtZero = 0; | 1201 | firstOccurrenceAtZero = 0; |
1196 | } | 1202 | } |
1197 | } | 1203 | } |
1198 | #endif | 1204 | #endif |
1199 | 1205 | ||
1200 | /* | 1206 | /* |
1201 | Tries to match in str and returns an array of (begin, length) pairs for | 1207 | Tries to match in str and returns an array of (begin, length) pairs for |
1202 | captured text. If there is no match, all pairs are (-1, -1). | 1208 | captured text. If there is no match, all pairs are (-1, -1). |
1203 | */ | 1209 | */ |
1204 | QArray<int> QRegExpEngine::match( const QString& str, int pos, bool minimal, | 1210 | QArray<int> QRegExpEngine::match( const QString& str, int pos, bool minimal, |
1205 | bool oneTest ) | 1211 | bool oneTest ) |
1206 | { | 1212 | { |
1207 | mmStr = &str; | 1213 | mmStr = &str; |
1208 | mmIn = str.unicode(); | 1214 | mmIn = str.unicode(); |
1209 | if ( mmIn == 0 ) | 1215 | if ( mmIn == 0 ) |
1210 | mmIn = &QChar::null; | 1216 | mmIn = &QChar::null; |
1211 | mmPos = pos; | 1217 | mmPos = pos; |
1212 | mmLen = str.length(); | 1218 | mmLen = str.length(); |
1213 | mmMinimal = minimal; | 1219 | mmMinimal = minimal; |
1214 | mmMatchedLen = 0; | 1220 | mmMatchedLen = 0; |
1215 | 1221 | ||
1216 | bool matched = FALSE; | 1222 | bool matched = FALSE; |
1217 | if ( valid && mmPos >= 0 && mmPos <= mmLen ) { | 1223 | if ( valid && mmPos >= 0 && mmPos <= mmLen ) { |
1218 | #ifndef QT_NO_REGEXP_OPTIM | 1224 | #ifndef QT_NO_REGEXP_OPTIM |
1219 | if ( mmPos <= mmLen - minl ) { | 1225 | if ( mmPos <= mmLen - minl ) { |
1220 | if ( caretAnchored || oneTest ) | 1226 | if ( caretAnchored || oneTest ) |
1221 | matched = matchHere(); | 1227 | matched = matchHere(); |
1222 | else if ( useGoodStringHeuristic ) | 1228 | else if ( useGoodStringHeuristic ) |
1223 | matched = goodStringMatch(); | 1229 | matched = goodStringMatch(); |
1224 | else | 1230 | else |
1225 | matched = badCharMatch(); | 1231 | matched = badCharMatch(); |
1226 | } | 1232 | } |
1227 | #else | 1233 | #else |
1228 | matched = oneTest ? matchHere() : bruteMatch(); | 1234 | matched = oneTest ? matchHere() : bruteMatch(); |
1229 | #endif | 1235 | #endif |
1230 | } | 1236 | } |
1231 | 1237 | ||
1232 | if ( matched ) { | 1238 | if ( matched ) { |
1233 | mmCaptured.detach(); | 1239 | mmCaptured.detach(); |
1234 | mmCaptured[0] = mmPos; | 1240 | mmCaptured[0] = mmPos; |
1235 | mmCaptured[1] = mmMatchedLen; | 1241 | mmCaptured[1] = mmMatchedLen; |
1236 | for ( int j = 0; j < realncap; j++ ) { | 1242 | for ( int j = 0; j < realncap; j++ ) { |
1237 | int len = mmCapEnd[j] - mmCapBegin[j]; | 1243 | int len = mmCapEnd[j] - mmCapBegin[j]; |
1238 | mmCaptured[2 + 2 * j] = len > 0 ? mmPos + mmCapBegin[j] : 0; | 1244 | mmCaptured[2 + 2 * j] = len > 0 ? mmPos + mmCapBegin[j] : 0; |
1239 | mmCaptured[2 + 2 * j + 1] = len; | 1245 | mmCaptured[2 + 2 * j + 1] = len; |
1240 | } | 1246 | } |
1241 | return mmCaptured; | 1247 | return mmCaptured; |
1242 | } else { | 1248 | } else { |
1243 | return mmCapturedNoMatch; | 1249 | return mmCapturedNoMatch; |
1244 | } | 1250 | } |
1245 | } | 1251 | } |
1246 | 1252 | ||
1247 | /* | 1253 | /* |
1248 | The three following functions add one state to the automaton and return the | 1254 | The three following functions add one state to the automaton and return the |
1249 | number of the state. | 1255 | number of the state. |
1250 | */ | 1256 | */ |
1251 | 1257 | ||
1252 | int QRegExpEngine::createState( QChar ch ) | 1258 | int QRegExpEngine::createState( QChar ch ) |
1253 | { | 1259 | { |
1254 | return setupState( ch.unicode() ); | 1260 | return setupState( ch.unicode() ); |
1255 | } | 1261 | } |
1256 | 1262 | ||
1257 | int QRegExpEngine::createState( const CharClass& cc ) | 1263 | int QRegExpEngine::createState( const CharClass& cc ) |
1258 | { | 1264 | { |
1259 | #ifndef QT_NO_REGEXP_CCLASS | 1265 | #ifndef QT_NO_REGEXP_CCLASS |
1260 | int n = cl.size(); | 1266 | int n = cl.size(); |
1261 | cl.resize( n + 1 ); | 1267 | cl.resize( n + 1 ); |
1262 | cl.insert( n, new CharClass(cc) ); | 1268 | cl.insert( n, new CharClass(cc) ); |
1263 | return setupState( CharClassBit | n ); | 1269 | return setupState( CharClassBit | n ); |
1264 | #else | 1270 | #else |
1265 | Q_UNUSED( cc ); | 1271 | Q_UNUSED( cc ); |
1266 | return setupState( CharClassBit ); | 1272 | return setupState( CharClassBit ); |
1267 | #endif | 1273 | #endif |
1268 | } | 1274 | } |
1269 | 1275 | ||
1270 | #ifndef QT_NO_REGEXP_BACKREF | 1276 | #ifndef QT_NO_REGEXP_BACKREF |
1271 | int QRegExpEngine::createState( int bref ) | 1277 | int QRegExpEngine::createState( int bref ) |
1272 | { | 1278 | { |
1273 | if ( bref > nbrefs ) { | 1279 | if ( bref > nbrefs ) { |
1274 | nbrefs = bref; | 1280 | nbrefs = bref; |
1275 | if ( nbrefs > MaxBackRefs ) { | 1281 | if ( nbrefs > MaxBackRefs ) { |
1276 | yyError = TRUE; | 1282 | yyError = TRUE; |
1277 | return 0; | 1283 | return 0; |
1278 | } | 1284 | } |
1279 | } | 1285 | } |
1280 | return setupState( BackRefBit | bref ); | 1286 | return setupState( BackRefBit | bref ); |
1281 | } | 1287 | } |
1282 | #endif | 1288 | #endif |
1283 | 1289 | ||
1284 | /* | 1290 | /* |
1285 | The two following functions add a transition between all pairs of states | 1291 | The two following functions add a transition between all pairs of states |
1286 | (i, j) where i is fond in from, and j is found in to. | 1292 | (i, j) where i is fond in from, and j is found in to. |
1287 | 1293 | ||
1288 | Cat-transitions are distinguished from plus-transitions for capturing. | 1294 | Cat-transitions are distinguished from plus-transitions for capturing. |
1289 | */ | 1295 | */ |
1290 | 1296 | ||
1291 | void QRegExpEngine::addCatTransitions( const QArray<int>& from, | 1297 | void QRegExpEngine::addCatTransitions( const QArray<int>& from, |
1292 | const QArray<int>& to ) | 1298 | const QArray<int>& to ) |
1293 | { | 1299 | { |
1294 | for ( int i = 0; i < (int) from.size(); i++ ) { | 1300 | for ( int i = 0; i < (int) from.size(); i++ ) { |
1295 | State *st = s[from[i]]; | 1301 | State *st = s[from[i]]; |
1296 | mergeInto( &st->outs, to ); | 1302 | mergeInto( &st->outs, to ); |
1297 | } | 1303 | } |
1298 | } | 1304 | } |
1299 | 1305 | ||
1300 | #ifndef QT_NO_REGEXP_CAPTURE | 1306 | #ifndef QT_NO_REGEXP_CAPTURE |
1301 | void QRegExpEngine::addPlusTransitions( const QArray<int>& from, | 1307 | void QRegExpEngine::addPlusTransitions( const QArray<int>& from, |
1302 | const QArray<int>& to, int atom ) | 1308 | const QArray<int>& to, int atom ) |
1303 | { | 1309 | { |
1304 | for ( int i = 0; i < (int) from.size(); i++ ) { | 1310 | for ( int i = 0; i < (int) from.size(); i++ ) { |
1305 | State *st = s[from[i]]; | 1311 | State *st = s[from[i]]; |
1306 | QArray<int> oldOuts = st->outs.copy(); | 1312 | QArray<int> oldOuts = st->outs.copy(); |
1307 | mergeInto( &st->outs, to ); | 1313 | mergeInto( &st->outs, to ); |
1308 | if ( f[atom].capture >= 0 ) { | 1314 | if ( f[atom].capture >= 0 ) { |
1309 | if ( st->reenter == 0 ) | 1315 | if ( st->reenter == 0 ) |
1310 | st->reenter = new QMap<int, int>; | 1316 | st->reenter = new QMap<int, int>; |
1311 | for ( int j = 0; j < (int) to.size(); j++ ) { | 1317 | for ( int j = 0; j < (int) to.size(); j++ ) { |
1312 | if ( !st->reenter->contains(to[j]) && | 1318 | if ( !st->reenter->contains(to[j]) && |
1313 | oldOuts.bsearch(to[j]) < 0 ) | 1319 | oldOuts.bsearch(to[j]) < 0 ) |
1314 | st->reenter->insert( to[j], atom ); | 1320 | st->reenter->insert( to[j], atom ); |
1315 | } | 1321 | } |
1316 | } | 1322 | } |
1317 | } | 1323 | } |
1318 | } | 1324 | } |
1319 | #endif | 1325 | #endif |
1320 | 1326 | ||
1321 | #ifndef QT_NO_REGEXP_ANCHOR_ALT | 1327 | #ifndef QT_NO_REGEXP_ANCHOR_ALT |
1322 | /* | 1328 | /* |
1323 | Returns an anchor that means a OR b. | 1329 | Returns an anchor that means a OR b. |
1324 | */ | 1330 | */ |
1325 | int QRegExpEngine::anchorAlternation( int a, int b ) | 1331 | int QRegExpEngine::anchorAlternation( int a, int b ) |
1326 | { | 1332 | { |
1327 | if ( ((a & b) == a || (a & b) == b) && ((a | b) & Anchor_Alternation) == 0 ) | 1333 | if ( ((a & b) == a || (a & b) == b) && ((a | b) & Anchor_Alternation) == 0 ) |
1328 | return a & b; | 1334 | return a & b; |
1329 | 1335 | ||
1330 | int n = aa.size(); | 1336 | int n = aa.size(); |
1331 | aa.resize( n + 1 ); | 1337 | aa.resize( n + 1 ); |
1332 | aa[n].a = a; | 1338 | aa[n].a = a; |
1333 | aa[n].b = b; | 1339 | aa[n].b = b; |
1334 | return Anchor_Alternation | n; | 1340 | return Anchor_Alternation | n; |
1335 | } | 1341 | } |
1336 | 1342 | ||
1337 | /* | 1343 | /* |
1338 | Returns an anchor that means a AND b. | 1344 | Returns an anchor that means a AND b. |
1339 | */ | 1345 | */ |
1340 | int QRegExpEngine::anchorConcatenation( int a, int b ) | 1346 | int QRegExpEngine::anchorConcatenation( int a, int b ) |
1341 | { | 1347 | { |
1342 | if ( ((a | b) & Anchor_Alternation) == 0 ) | 1348 | if ( ((a | b) & Anchor_Alternation) == 0 ) |
1343 | return a | b; | 1349 | return a | b; |
1344 | if ( (b & Anchor_Alternation) != 0 ) | 1350 | if ( (b & Anchor_Alternation) != 0 ) |
1345 | qSwap( a, b ); | 1351 | qSwap( a, b ); |
1346 | int aprime = anchorConcatenation( aa[a ^ Anchor_Alternation].a, b ); | 1352 | int aprime = anchorConcatenation( aa[a ^ Anchor_Alternation].a, b ); |
1347 | int bprime = anchorConcatenation( aa[a ^ Anchor_Alternation].b, b ); | 1353 | int bprime = anchorConcatenation( aa[a ^ Anchor_Alternation].b, b ); |
1348 | return anchorAlternation( aprime, bprime ); | 1354 | return anchorAlternation( aprime, bprime ); |
1349 | } | 1355 | } |
1350 | #endif | 1356 | #endif |
1351 | 1357 | ||
1352 | /* | 1358 | /* |
1353 | Adds anchor a on a transition caracterised by its from state and its to state. | 1359 | Adds anchor a on a transition caracterised by its from state and its to state. |
1354 | */ | 1360 | */ |
1355 | void QRegExpEngine::addAnchors( int from, int to, int a ) | 1361 | void QRegExpEngine::addAnchors( int from, int to, int a ) |
1356 | { | 1362 | { |
1357 | State *st = s[from]; | 1363 | State *st = s[from]; |
1358 | if ( st->anchors == 0 ) | 1364 | if ( st->anchors == 0 ) |
1359 | st->anchors = new QMap<int, int>; | 1365 | st->anchors = new QMap<int, int>; |
1360 | if ( st->anchors->contains(to) ) | 1366 | if ( st->anchors->contains(to) ) |
1361 | a = anchorAlternation( (*st->anchors)[to], a ); | 1367 | a = anchorAlternation( (*st->anchors)[to], a ); |
1362 | st->anchors->insert( to, a ); | 1368 | st->anchors->insert( to, a ); |
1363 | } | 1369 | } |
1364 | 1370 | ||
1365 | #ifndef QT_NO_REGEXP_OPTIM | 1371 | #ifndef QT_NO_REGEXP_OPTIM |
1366 | /* | 1372 | /* |
1367 | The two following functions provide the engine with the information needed by | 1373 | The two following functions provide the engine with the information needed by |
1368 | its matching heuristics. | 1374 | its matching heuristics. |
1369 | */ | 1375 | */ |
1370 | 1376 | ||
1371 | void QRegExpEngine::setupGoodStringHeuristic( int earlyStart, int lateStart, | 1377 | void QRegExpEngine::setupGoodStringHeuristic( int earlyStart, int lateStart, |
1372 | const QString& str ) | 1378 | const QString& str ) |
1373 | { | 1379 | { |
1374 | goodEarlyStart = earlyStart; | 1380 | goodEarlyStart = earlyStart; |
1375 | goodLateStart = lateStart; | 1381 | goodLateStart = lateStart; |
1376 | goodStr = cs ? str : str.lower(); | 1382 | goodStr = cs ? str : str.lower(); |
1377 | } | 1383 | } |
1378 | 1384 | ||
1379 | void QRegExpEngine::setupBadCharHeuristic( int minLen, | 1385 | void QRegExpEngine::setupBadCharHeuristic( int minLen, |
1380 | const QArray<int>& firstOcc ) | 1386 | const QArray<int>& firstOcc ) |
1381 | { | 1387 | { |
1382 | minl = minLen; | 1388 | minl = minLen; |
1383 | occ1 = cs ? firstOcc : *firstOccurrenceAtZero; | 1389 | occ1 = cs ? firstOcc : *firstOccurrenceAtZero; |
1384 | } | 1390 | } |
1385 | 1391 | ||
1386 | /* | 1392 | /* |
1387 | This function chooses between the good-string and the bad-character | 1393 | This function chooses between the good-string and the bad-character |
1388 | heuristics. It computes two scores and chooses the heuristic with the highest | 1394 | heuristics. It computes two scores and chooses the heuristic with the highest |
1389 | score. | 1395 | score. |
1390 | 1396 | ||
1391 | Here are some common-sense constraints on the scores that should be respected | 1397 | Here are some common-sense constraints on the scores that should be respected |
1392 | if the formulas are ever modified: (1) If goodStr is empty, the good-string | 1398 | if the formulas are ever modified: (1) If goodStr is empty, the good-string |
1393 | heuristic scores 0. (2) If the search is case insensitive, the good-string | 1399 | heuristic scores 0. (2) If the search is case insensitive, the good-string |
1394 | heuristic should be used, unless it scores 0. (Case insensitivity | 1400 | heuristic should be used, unless it scores 0. (Case insensitivity |
1395 | turns all entries of occ1 to 0.) (3) If (goodLateStart - goodEarlyStart) is | 1401 | turns all entries of occ1 to 0.) (3) If (goodLateStart - goodEarlyStart) is |
1396 | big, the good-string heuristic should score less. | 1402 | big, the good-string heuristic should score less. |
1397 | */ | 1403 | */ |
1398 | void QRegExpEngine::heuristicallyChooseHeuristic() | 1404 | void QRegExpEngine::heuristicallyChooseHeuristic() |
1399 | { | 1405 | { |
1400 | int i; | 1406 | int i; |
1401 | 1407 | ||
1402 | if ( minl == 0 ) | 1408 | if ( minl == 0 ) |
1403 | return; | 1409 | return; |
1404 | 1410 | ||
1405 | /* | 1411 | /* |
1406 | Magic formula: The good string has to constitute a good proportion of the | 1412 | Magic formula: The good string has to constitute a good proportion of the |
1407 | minimum-length string, and appear at a more-or-less known index. | 1413 | minimum-length string, and appear at a more-or-less known index. |
1408 | */ | 1414 | */ |
1409 | int goodStringScore = ( 64 * goodStr.length() / minl ) - | 1415 | int goodStringScore = ( 64 * goodStr.length() / minl ) - |
1410 | ( goodLateStart - goodEarlyStart ); | 1416 | ( goodLateStart - goodEarlyStart ); |
1411 | 1417 | ||
1412 | /* | 1418 | /* |
1413 | Less magic formula: We pick a couple of characters at random, and check | 1419 | Less magic formula: We pick a couple of characters at random, and check |
1414 | whether they are good or bad. | 1420 | whether they are good or bad. |
1415 | */ | 1421 | */ |
1416 | int badCharScore = 0; | 1422 | int badCharScore = 0; |
1417 | int step = QMAX( 1, NumBadChars / 32 ); | 1423 | int step = QMAX( 1, NumBadChars / 32 ); |
1418 | for ( i = 1; i < NumBadChars; i += step ) { | 1424 | for ( i = 1; i < NumBadChars; i += step ) { |
1419 | if ( occ1[i] == NoOccurrence ) | 1425 | if ( occ1[i] == NoOccurrence ) |
1420 | badCharScore += minl; | 1426 | badCharScore += minl; |
1421 | else | 1427 | else |
1422 | badCharScore += occ1[i]; | 1428 | badCharScore += occ1[i]; |
1423 | } | 1429 | } |
1424 | badCharScore /= minl; | 1430 | badCharScore /= minl; |
1425 | 1431 | ||
1426 | useGoodStringHeuristic = ( goodStringScore > badCharScore ); | 1432 | useGoodStringHeuristic = ( goodStringScore > badCharScore ); |
1427 | } | 1433 | } |
1428 | #endif | 1434 | #endif |
1429 | 1435 | ||
1430 | #if defined(QT_DEBUG) | 1436 | #if defined(QT_DEBUG) |
1431 | void QRegExpEngine::dump() const | 1437 | void QRegExpEngine::dump() const |
1432 | { | 1438 | { |
1433 | int i, j; | 1439 | int i, j; |
1434 | qDebug( "Case %ssensitive engine", cs ? "" : "in" ); | 1440 | odebug << "Case " << (cs ? "" : "in") << "sensitive engine" << oendl; |
1435 | qDebug( " States" ); | 1441 | odebug << " States" << oendl; |
1436 | for ( i = 0; i < ns; i++ ) { | 1442 | for ( i = 0; i < ns; i++ ) { |
1437 | qDebug( " %d%s", i, | 1443 | odebug << " " << i |
1438 | i == InitialState ? " (initial)" : | 1444 | << (i == InitialState ? " (initial)" : i == FinalState ? " (final)" : "") << oendl; |
1439 | i == FinalState ? " (final)" : "" ); | 1445 | |
1440 | #ifndef QT_NO_REGEXP_CAPTURE | 1446 | #ifndef QT_NO_REGEXP_CAPTURE |
1441 | qDebug( " in atom %d", s[i]->atom ); | 1447 | odebug << " in atom " << s[i]->atom << oendl; |
1442 | #endif | 1448 | #endif |
1443 | int m = s[i]->match; | 1449 | int m = s[i]->match; |
1444 | if ( (m & CharClassBit) != 0 ) { | 1450 | if ( (m & CharClassBit) != 0 ) { |
1445 | qDebug( " match character class %d", m ^ CharClassBit ); | 1451 | odebug << " match character class " << (m ^ CharClassBit) << oendl; |
1446 | #ifndef QT_NO_REGEXP_CCLASS | 1452 | #ifndef QT_NO_REGEXP_CCLASS |
1447 | cl[m ^ CharClassBit]->dump(); | 1453 | cl[m ^ CharClassBit]->dump(); |
1448 | #else | 1454 | #else |
1449 | qDebug( " negative character class" ); | 1455 | odebug << " negative character class" << oendl; |
1450 | #endif | 1456 | #endif |
1451 | } else if ( (m & BackRefBit) != 0 ) { | 1457 | } else if ( (m & BackRefBit) != 0 ) { |
1452 | qDebug( " match back-reference %d", m ^ BackRefBit ); | 1458 | odebug << " match back-reference " << (m ^ BackRefBit) << oendl; |
1453 | } else if ( m >= 0x20 && m <= 0x7e ) { | 1459 | } else if ( m >= 0x20 && m <= 0x7e ) { |
1454 | qDebug( " match 0x%.4x (%c)", m, m ); | 1460 | odebug << " match " << QString().sprintf( "0x%.4x", m) << " (" << m << ")" << oendl; |
1455 | } else { | 1461 | |
1456 | qDebug( " match 0x%.4x", m ); | 1462 | } else { |
1457 | } | 1463 | odebug << " match " << QString().sprintf( "0x%.4x", m) << oendl; |
1458 | for ( j = 0; j < (int) s[i]->outs.size(); j++ ) { | 1464 | } |
1459 | int next = s[i]->outs[j]; | 1465 | for ( j = 0; j < (int) s[i]->outs.size(); j++ ) { |
1460 | qDebug( " -> %d", next ); | 1466 | int next = s[i]->outs[j]; |
1461 | if ( s[i]->reenter != 0 && s[i]->reenter->contains(next) ) | 1467 | odebug << " -> " << next << oendl; |
1462 | qDebug( " [reenter %d]", (*s[i]->reenter)[next] ); | 1468 | if ( s[i]->reenter != 0 && s[i]->reenter->contains(next) ) |
1463 | if ( s[i]->anchors != 0 && at(*s[i]->anchors, next) != 0 ) | 1469 | odebug << " [reenter " << (*s[i]->reenter)[next] << "]" << oendl; |
1464 | qDebug( " [anchors 0x%.8x]", (*s[i]->anchors)[next] ); | 1470 | if ( s[i]->anchors != 0 && at(*s[i]->anchors, next) != 0 ) |
1465 | } | 1471 | odebug << " [anchors " << QString().sprintf( "0x%.8x]", (*s[i]->anchors)[next] ) << oendl; |
1472 | } | ||
1466 | } | 1473 | } |
1467 | #ifndef QT_NO_REGEXP_CAPTURE | 1474 | #ifndef QT_NO_REGEXP_CAPTURE |
1468 | if ( nf > 0 ) { | 1475 | if ( nf > 0 ) { |
1469 | qDebug( " Atom Parent Capture" ); | 1476 | odebug << " Atom Parent Capture" << oendl; |
1470 | for ( i = 0; i < nf; i++ ) | 1477 | for ( i = 0; i < nf; i++ ) |
1471 | qDebug( " %6d %6d %6d", i, f[i].parent, f[i].capture ); | 1478 | odebug << QString().sprintf(" %6d %6d %6d", i, f[i].parent, f[i].capture ) << oendl; |
1472 | } | 1479 | } |
1473 | #endif | 1480 | #endif |
1474 | #ifndef QT_NO_REGEXP_ANCHOR_ALT | 1481 | #ifndef QT_NO_REGEXP_ANCHOR_ALT |
1475 | for ( i = 0; i < (int) aa.size(); i++ ) | 1482 | for ( i = 0; i < (int) aa.size(); i++ ) |
1476 | qDebug( " Anchor alternation 0x%.8x: 0x%.8x 0x%.9x", i, aa[i].a, | 1483 | odebug << QString().sprintf(" Anchor alternation 0x%.8x: 0x%.8x 0x%.9x", i, aa[i].a, aa[i].b ) << oendl; |
1477 | aa[i].b ); | ||
1478 | #endif | 1484 | #endif |
1479 | } | 1485 | } |
1480 | #endif | 1486 | #endif |
1481 | 1487 | ||
1482 | void QRegExpEngine::setup( bool caseSensitive ) | 1488 | void QRegExpEngine::setup( bool caseSensitive ) |
1483 | { | 1489 | { |
1484 | #ifndef QT_NO_REGEXP_OPTIM | 1490 | #ifndef QT_NO_REGEXP_OPTIM |
1485 | if ( engCount++ == 0 ) { | 1491 | if ( engCount++ == 0 ) { |
1486 | noOccurrences = new QArray<int>( NumBadChars ); | 1492 | noOccurrences = new QArray<int>( NumBadChars ); |
1487 | firstOccurrenceAtZero = new QArray<int>( NumBadChars ); | 1493 | firstOccurrenceAtZero = new QArray<int>( NumBadChars ); |
1488 | noOccurrences->fill( NoOccurrence ); | 1494 | noOccurrences->fill( NoOccurrence ); |
1489 | firstOccurrenceAtZero->fill( 0 ); | 1495 | firstOccurrenceAtZero->fill( 0 ); |
1490 | } | 1496 | } |
1491 | #endif | 1497 | #endif |
1492 | s.setAutoDelete( TRUE ); | 1498 | s.setAutoDelete( TRUE ); |
1493 | s.resize( 32 ); | 1499 | s.resize( 32 ); |
1494 | ns = 0; | 1500 | ns = 0; |
1495 | #ifndef QT_NO_REGEXP_CAPTURE | 1501 | #ifndef QT_NO_REGEXP_CAPTURE |
1496 | f.resize( 32 ); | 1502 | f.resize( 32 ); |
1497 | nf = 0; | 1503 | nf = 0; |
1498 | cf = -1; | 1504 | cf = -1; |
1499 | #endif | 1505 | #endif |
1500 | realncap = 0; | 1506 | realncap = 0; |
1501 | ncap = 0; | 1507 | ncap = 0; |
1502 | #ifndef QT_NO_REGEXP_CCLASS | 1508 | #ifndef QT_NO_REGEXP_CCLASS |
1503 | cl.setAutoDelete( TRUE ); | 1509 | cl.setAutoDelete( TRUE ); |
1504 | #endif | 1510 | #endif |
1505 | #ifndef QT_NO_REGEXP_LOOKAHEAD | 1511 | #ifndef QT_NO_REGEXP_LOOKAHEAD |
1506 | ahead.setAutoDelete( TRUE ); | 1512 | ahead.setAutoDelete( TRUE ); |
1507 | #endif | 1513 | #endif |
1508 | #ifndef QT_NO_REGEXP_OPTIM | 1514 | #ifndef QT_NO_REGEXP_OPTIM |
1509 | caretAnchored = TRUE; | 1515 | caretAnchored = TRUE; |
1510 | #endif | 1516 | #endif |
1511 | valid = FALSE; | 1517 | valid = FALSE; |
1512 | cs = caseSensitive; | 1518 | cs = caseSensitive; |
1513 | #ifndef QT_NO_REGEXP_BACKREF | 1519 | #ifndef QT_NO_REGEXP_BACKREF |
1514 | nbrefs = 0; | 1520 | nbrefs = 0; |
1515 | #endif | 1521 | #endif |
1516 | #ifndef QT_NO_REGEXP_OPTIM | 1522 | #ifndef QT_NO_REGEXP_OPTIM |
1517 | useGoodStringHeuristic = FALSE; | 1523 | useGoodStringHeuristic = FALSE; |
1518 | minl = 0; | 1524 | minl = 0; |
1519 | occ1 = *firstOccurrenceAtZero; | 1525 | occ1 = *firstOccurrenceAtZero; |
1520 | #endif | 1526 | #endif |
1521 | mmCapturedNoMatch.fill( -1, 2 ); | 1527 | mmCapturedNoMatch.fill( -1, 2 ); |
1522 | } | 1528 | } |
1523 | 1529 | ||
1524 | int QRegExpEngine::setupState( int match ) | 1530 | int QRegExpEngine::setupState( int match ) |
1525 | { | 1531 | { |
1526 | if ( (ns & (ns + 1)) == 0 && ns + 1 >= (int) s.size() ) | 1532 | if ( (ns & (ns + 1)) == 0 && ns + 1 >= (int) s.size() ) |
1527 | s.resize( (ns + 1) << 1 ); | 1533 | s.resize( (ns + 1) << 1 ); |
1528 | #ifndef QT_NO_REGEXP_CAPTURE | 1534 | #ifndef QT_NO_REGEXP_CAPTURE |
1529 | s.insert( ns, new State(cf, match) ); | 1535 | s.insert( ns, new State(cf, match) ); |
1530 | #else | 1536 | #else |
1531 | s.insert( ns, new State(match) ); | 1537 | s.insert( ns, new State(match) ); |
1532 | #endif | 1538 | #endif |
1533 | return ns++; | 1539 | return ns++; |
1534 | } | 1540 | } |
1535 | 1541 | ||
1536 | #ifndef QT_NO_REGEXP_CAPTURE | 1542 | #ifndef QT_NO_REGEXP_CAPTURE |
1537 | /* | 1543 | /* |
1538 | Functions startAtom() and finishAtom() should be called to delimit atoms. | 1544 | Functions startAtom() and finishAtom() should be called to delimit atoms. |
1539 | When a state is created, it is assigned to the current atom. The information | 1545 | When a state is created, it is assigned to the current atom. The information |
1540 | is later used for capturing. | 1546 | is later used for capturing. |
1541 | */ | 1547 | */ |
1542 | int QRegExpEngine::startAtom( bool capture ) | 1548 | int QRegExpEngine::startAtom( bool capture ) |
1543 | { | 1549 | { |
1544 | if ( (nf & (nf + 1)) == 0 && nf + 1 >= (int) f.size() ) | 1550 | if ( (nf & (nf + 1)) == 0 && nf + 1 >= (int) f.size() ) |
1545 | f.resize( (nf + 1) << 1 ); | 1551 | f.resize( (nf + 1) << 1 ); |
1546 | f[nf].parent = cf; | 1552 | f[nf].parent = cf; |
1547 | cf = nf++; | 1553 | cf = nf++; |
1548 | f[cf].capture = capture ? ncap++ : -1; | 1554 | f[cf].capture = capture ? ncap++ : -1; |
1549 | return cf; | 1555 | return cf; |
1550 | } | 1556 | } |
1551 | #endif | 1557 | #endif |
1552 | 1558 | ||
1553 | #ifndef QT_NO_REGEXP_LOOKAHEAD | 1559 | #ifndef QT_NO_REGEXP_LOOKAHEAD |
1554 | /* | 1560 | /* |
1555 | Creates a lookahead anchor. | 1561 | Creates a lookahead anchor. |
1556 | */ | 1562 | */ |
1557 | int QRegExpEngine::addLookahead( QRegExpEngine *eng, bool negative ) | 1563 | int QRegExpEngine::addLookahead( QRegExpEngine *eng, bool negative ) |
1558 | { | 1564 | { |
1559 | int n = ahead.size(); | 1565 | int n = ahead.size(); |
1560 | if ( n == MaxLookaheads ) { | 1566 | if ( n == MaxLookaheads ) { |
1561 | yyError = TRUE; | 1567 | yyError = TRUE; |
1562 | return 0; | 1568 | return 0; |
1563 | } | 1569 | } |
1564 | ahead.resize( n + 1 ); | 1570 | ahead.resize( n + 1 ); |
1565 | ahead.insert( n, new Lookahead(eng, negative) ); | 1571 | ahead.insert( n, new Lookahead(eng, negative) ); |
1566 | return Anchor_FirstLookahead << n; | 1572 | return Anchor_FirstLookahead << n; |
1567 | } | 1573 | } |
1568 | #endif | 1574 | #endif |
1569 | 1575 | ||
1570 | #ifndef QT_NO_REGEXP_CAPTURE | 1576 | #ifndef QT_NO_REGEXP_CAPTURE |
1571 | /* | 1577 | /* |
1572 | We want the longest leftmost captures. | 1578 | We want the longest leftmost captures. |
1573 | */ | 1579 | */ |
1574 | bool QRegExpEngine::isBetterCapture( const int *begin1, const int *end1, | 1580 | bool QRegExpEngine::isBetterCapture( const int *begin1, const int *end1, |
1575 | const int *begin2, const int *end2 ) | 1581 | const int *begin2, const int *end2 ) |
1576 | { | 1582 | { |
1577 | for ( int i = 0; i < ncap; i++ ) { | 1583 | for ( int i = 0; i < ncap; i++ ) { |
1578 | int delta = begin2[i] - begin1[i]; // it has to start early... | 1584 | int delta = begin2[i] - begin1[i]; // it has to start early... |
1579 | if ( delta == 0 ) | 1585 | if ( delta == 0 ) |
1580 | delta = end1[i] - end2[i]; // ...and end late (like a party) | 1586 | delta = end1[i] - end2[i]; // ...and end late (like a party) |
1581 | 1587 | ||
1582 | if ( delta != 0 ) | 1588 | if ( delta != 0 ) |
1583 | return delta > 0; | 1589 | return delta > 0; |
1584 | } | 1590 | } |
1585 | return FALSE; | 1591 | return FALSE; |
1586 | } | 1592 | } |
1587 | #endif | 1593 | #endif |
1588 | 1594 | ||
1589 | /* | 1595 | /* |
1590 | Returns TRUE if anchor a matches at position mmPos + i in the input string, | 1596 | Returns TRUE if anchor a matches at position mmPos + i in the input string, |
1591 | otherwise FALSE. | 1597 | otherwise FALSE. |
1592 | */ | 1598 | */ |
1593 | bool QRegExpEngine::testAnchor( int i, int a, const int *capBegin ) | 1599 | bool QRegExpEngine::testAnchor( int i, int a, const int *capBegin ) |
1594 | { | 1600 | { |
1595 | int j; | 1601 | int j; |
1596 | 1602 | ||
1597 | #ifndef QT_NO_REGEXP_ANCHOR_ALT | 1603 | #ifndef QT_NO_REGEXP_ANCHOR_ALT |
1598 | if ( (a & Anchor_Alternation) != 0 ) { | 1604 | if ( (a & Anchor_Alternation) != 0 ) { |
1599 | return testAnchor( i, aa[a ^ Anchor_Alternation].a, capBegin ) || | 1605 | return testAnchor( i, aa[a ^ Anchor_Alternation].a, capBegin ) || |
1600 | testAnchor( i, aa[a ^ Anchor_Alternation].b, capBegin ); | 1606 | testAnchor( i, aa[a ^ Anchor_Alternation].b, capBegin ); |
1601 | } | 1607 | } |
1602 | #endif | 1608 | #endif |
1603 | 1609 | ||
1604 | if ( (a & Anchor_Caret) != 0 ) { | 1610 | if ( (a & Anchor_Caret) != 0 ) { |
1605 | if ( mmPos + i != 0 ) | 1611 | if ( mmPos + i != 0 ) |
1606 | return FALSE; | 1612 | return FALSE; |
1607 | } | 1613 | } |
1608 | if ( (a & Anchor_Dollar) != 0 ) { | 1614 | if ( (a & Anchor_Dollar) != 0 ) { |
1609 | if ( mmPos + i != mmLen ) | 1615 | if ( mmPos + i != mmLen ) |
1610 | return FALSE; | 1616 | return FALSE; |
1611 | } | 1617 | } |
1612 | #ifndef QT_NO_REGEXP_ESCAPE | 1618 | #ifndef QT_NO_REGEXP_ESCAPE |
1613 | if ( (a & (Anchor_Word | Anchor_NonWord)) != 0 ) { | 1619 | if ( (a & (Anchor_Word | Anchor_NonWord)) != 0 ) { |
1614 | bool before = FALSE, after = FALSE; | 1620 | bool before = FALSE, after = FALSE; |
1615 | if ( mmPos + i != 0 ) | 1621 | if ( mmPos + i != 0 ) |
1616 | before = mmIn[mmPos + i - 1].isLetterOrNumber(); | 1622 | before = mmIn[mmPos + i - 1].isLetterOrNumber(); |
1617 | if ( mmPos + i != mmLen ) | 1623 | if ( mmPos + i != mmLen ) |
1618 | after = mmIn[mmPos + i].isLetterOrNumber(); | 1624 | after = mmIn[mmPos + i].isLetterOrNumber(); |
1619 | if ( (a & Anchor_Word) != 0 && (before == after) ) | 1625 | if ( (a & Anchor_Word) != 0 && (before == after) ) |
1620 | return FALSE; | 1626 | return FALSE; |
1621 | if ( (a & Anchor_NonWord) != 0 && (before != after) ) | 1627 | if ( (a & Anchor_NonWord) != 0 && (before != after) ) |
1622 | return FALSE; | 1628 | return FALSE; |
1623 | } | 1629 | } |
1624 | #endif | 1630 | #endif |
1625 | #ifndef QT_NO_REGEXP_LOOKAHEAD | 1631 | #ifndef QT_NO_REGEXP_LOOKAHEAD |
1626 | bool catchx = TRUE; | 1632 | bool catchx = TRUE; |
1627 | 1633 | ||
1628 | if ( (a & Anchor_LookaheadMask) != 0 ) { | 1634 | if ( (a & Anchor_LookaheadMask) != 0 ) { |
1629 | QConstString cstr = QConstString( (QChar *) mmIn + mmPos + i, | 1635 | QConstString cstr = QConstString( (QChar *) mmIn + mmPos + i, |
1630 | mmLen - mmPos - i ); | 1636 | mmLen - mmPos - i ); |
1631 | for ( j = 0; j < (int) ahead.size(); j++ ) { | 1637 | for ( j = 0; j < (int) ahead.size(); j++ ) { |
1632 | if ( (a & (Anchor_FirstLookahead << j)) != 0 ) { | 1638 | if ( (a & (Anchor_FirstLookahead << j)) != 0 ) { |
1633 | catchx = ( ahead[j]->eng->match(cstr.string(), 0, TRUE, | 1639 | catchx = ( ahead[j]->eng->match(cstr.string(), 0, TRUE, |
1634 | TRUE)[0] == 0 ); | 1640 | TRUE)[0] == 0 ); |
1635 | if ( catchx == ahead[j]->neg ) | 1641 | if ( catchx == ahead[j]->neg ) |
1636 | return FALSE; | 1642 | return FALSE; |
1637 | } | 1643 | } |
1638 | } | 1644 | } |
1639 | } | 1645 | } |
1640 | #endif | 1646 | #endif |
1641 | #ifndef QT_NO_REGEXP_CAPTURE | 1647 | #ifndef QT_NO_REGEXP_CAPTURE |
1642 | #ifndef QT_NO_REGEXP_BACKREF | 1648 | #ifndef QT_NO_REGEXP_BACKREF |
1643 | for ( j = 0; j < nbrefs; j++ ) { | 1649 | for ( j = 0; j < nbrefs; j++ ) { |
1644 | if ( (a & (Anchor_BackRef1Empty << j)) != 0 ) { | 1650 | if ( (a & (Anchor_BackRef1Empty << j)) != 0 ) { |
1645 | if ( capBegin[j] != EmptyCapture ) | 1651 | if ( capBegin[j] != EmptyCapture ) |
1646 | return FALSE; | 1652 | return FALSE; |
1647 | } | 1653 | } |
1648 | } | 1654 | } |
1649 | #endif | 1655 | #endif |
1650 | #endif | 1656 | #endif |
1651 | return TRUE; | 1657 | return TRUE; |
1652 | } | 1658 | } |
1653 | 1659 | ||
1654 | #ifndef QT_NO_REGEXP_OPTIM | 1660 | #ifndef QT_NO_REGEXP_OPTIM |
1655 | /* | 1661 | /* |
1656 | The three following functions are what Jeffrey Friedl would call transmissions | 1662 | The three following functions are what Jeffrey Friedl would call transmissions |
1657 | (or bump-alongs). Using one or the other should make no difference, except | 1663 | (or bump-alongs). Using one or the other should make no difference, except |
1658 | in performance. | 1664 | in performance. |
1659 | */ | 1665 | */ |
1660 | 1666 | ||
1661 | bool QRegExpEngine::goodStringMatch() | 1667 | bool QRegExpEngine::goodStringMatch() |
1662 | { | 1668 | { |
1663 | int k = mmPos + goodEarlyStart; | 1669 | int k = mmPos + goodEarlyStart; |
1664 | 1670 | ||
1665 | while ( (k = mmStr->find(goodStr, k, cs)) != -1 ) { | 1671 | while ( (k = mmStr->find(goodStr, k, cs)) != -1 ) { |
1666 | int from = k - goodLateStart; | 1672 | int from = k - goodLateStart; |
1667 | int to = k - goodEarlyStart; | 1673 | int to = k - goodEarlyStart; |
1668 | if ( from > mmPos ) | 1674 | if ( from > mmPos ) |
1669 | mmPos = from; | 1675 | mmPos = from; |
1670 | 1676 | ||
1671 | while ( mmPos <= to ) { | 1677 | while ( mmPos <= to ) { |
1672 | if ( matchHere() ) | 1678 | if ( matchHere() ) |
1673 | return TRUE; | 1679 | return TRUE; |
1674 | mmPos++; | 1680 | mmPos++; |
1675 | } | 1681 | } |
1676 | k++; | 1682 | k++; |
1677 | } | 1683 | } |
1678 | return FALSE; | 1684 | return FALSE; |
1679 | } | 1685 | } |
1680 | 1686 | ||
1681 | bool QRegExpEngine::badCharMatch() | 1687 | bool QRegExpEngine::badCharMatch() |
1682 | { | 1688 | { |
1683 | int slideHead = 0; | 1689 | int slideHead = 0; |
1684 | int slideNext = 0; | 1690 | int slideNext = 0; |
1685 | int i; | 1691 | int i; |
1686 | int lastPos = mmLen - minl; | 1692 | int lastPos = mmLen - minl; |
1687 | memset( mmSlideTab, 0, mmSlideTabSize * sizeof(int) ); | 1693 | memset( mmSlideTab, 0, mmSlideTabSize * sizeof(int) ); |
1688 | 1694 | ||
1689 | /* | 1695 | /* |
1690 | Set up the slide table, used for the bad-character heuristic, using | 1696 | Set up the slide table, used for the bad-character heuristic, using |
1691 | the table of first occurrence of each character. | 1697 | the table of first occurrence of each character. |
1692 | */ | 1698 | */ |
1693 | for ( i = 0; i < minl; i++ ) { | 1699 | for ( i = 0; i < minl; i++ ) { |
1694 | int sk = occ1[BadChar(mmIn[mmPos + i])]; | 1700 | int sk = occ1[BadChar(mmIn[mmPos + i])]; |
1695 | if ( sk == NoOccurrence ) | 1701 | if ( sk == NoOccurrence ) |
1696 | sk = i + 1; | 1702 | sk = i + 1; |
1697 | if ( sk > 0 ) { | 1703 | if ( sk > 0 ) { |
1698 | int k = i + 1 - sk; | 1704 | int k = i + 1 - sk; |
1699 | if ( k < 0 ) { | 1705 | if ( k < 0 ) { |
1700 | sk = i + 1; | 1706 | sk = i + 1; |
1701 | k = 0; | 1707 | k = 0; |
1702 | } | 1708 | } |
1703 | if ( sk > mmSlideTab[k] ) | 1709 | if ( sk > mmSlideTab[k] ) |
1704 | mmSlideTab[k] = sk; | 1710 | mmSlideTab[k] = sk; |
1705 | } | 1711 | } |
1706 | } | 1712 | } |
1707 | 1713 | ||
1708 | if ( mmPos > lastPos ) | 1714 | if ( mmPos > lastPos ) |
1709 | return FALSE; | 1715 | return FALSE; |
1710 | 1716 | ||
1711 | while ( TRUE ) { | 1717 | while ( TRUE ) { |
1712 | if ( ++slideNext >= mmSlideTabSize ) | 1718 | if ( ++slideNext >= mmSlideTabSize ) |
1713 | slideNext = 0; | 1719 | slideNext = 0; |
1714 | if ( mmSlideTab[slideHead] > 0 ) { | 1720 | if ( mmSlideTab[slideHead] > 0 ) { |
1715 | if ( mmSlideTab[slideHead] - 1 > mmSlideTab[slideNext] ) | 1721 | if ( mmSlideTab[slideHead] - 1 > mmSlideTab[slideNext] ) |
1716 | mmSlideTab[slideNext] = mmSlideTab[slideHead] - 1; | 1722 | mmSlideTab[slideNext] = mmSlideTab[slideHead] - 1; |
1717 | mmSlideTab[slideHead] = 0; | 1723 | mmSlideTab[slideHead] = 0; |
1718 | } else { | 1724 | } else { |
1719 | if ( matchHere() ) | 1725 | if ( matchHere() ) |
1720 | return TRUE; | 1726 | return TRUE; |
1721 | } | 1727 | } |
1722 | 1728 | ||
1723 | if ( mmPos == lastPos ) | 1729 | if ( mmPos == lastPos ) |
1724 | break; | 1730 | break; |
1725 | 1731 | ||
1726 | /* | 1732 | /* |
1727 | Update the slide table. This code has much in common with the | 1733 | Update the slide table. This code has much in common with the |
1728 | initialization code. | 1734 | initialization code. |
1729 | */ | 1735 | */ |
1730 | int sk = occ1[BadChar(mmIn[mmPos + minl])]; | 1736 | int sk = occ1[BadChar(mmIn[mmPos + minl])]; |
1731 | if ( sk == NoOccurrence ) { | 1737 | if ( sk == NoOccurrence ) { |
1732 | mmSlideTab[slideNext] = minl; | 1738 | mmSlideTab[slideNext] = minl; |
1733 | } else if ( sk > 0 ) { | 1739 | } else if ( sk > 0 ) { |
1734 | int k = slideNext + minl - sk; | 1740 | int k = slideNext + minl - sk; |
1735 | if ( k >= mmSlideTabSize ) | 1741 | if ( k >= mmSlideTabSize ) |
1736 | k -= mmSlideTabSize; | 1742 | k -= mmSlideTabSize; |
1737 | if ( sk > mmSlideTab[k] ) | 1743 | if ( sk > mmSlideTab[k] ) |
1738 | mmSlideTab[k] = sk; | 1744 | mmSlideTab[k] = sk; |
1739 | } | 1745 | } |
1740 | slideHead = slideNext; | 1746 | slideHead = slideNext; |
1741 | mmPos++; | 1747 | mmPos++; |
1742 | } | 1748 | } |
1743 | return FALSE; | 1749 | return FALSE; |
1744 | } | 1750 | } |
1745 | #else | 1751 | #else |
1746 | bool QRegExpEngine::bruteMatch() | 1752 | bool QRegExpEngine::bruteMatch() |
1747 | { | 1753 | { |
1748 | while ( mmPos <= mmLen ) { | 1754 | while ( mmPos <= mmLen ) { |
1749 | if ( matchHere() ) | 1755 | if ( matchHere() ) |
1750 | return TRUE; | 1756 | return TRUE; |
1751 | mmPos++; | 1757 | mmPos++; |
1752 | } | 1758 | } |
1753 | return FALSE; | 1759 | return FALSE; |
1754 | } | 1760 | } |
1755 | #endif | 1761 | #endif |
1756 | 1762 | ||
1757 | /* | 1763 | /* |
1758 | Here's the core of the engine. It tries to do a match here and now. | 1764 | Here's the core of the engine. It tries to do a match here and now. |
1759 | */ | 1765 | */ |
1760 | bool QRegExpEngine::matchHere() | 1766 | bool QRegExpEngine::matchHere() |
1761 | { | 1767 | { |
1762 | int ncur = 1, nnext = 0; | 1768 | int ncur = 1, nnext = 0; |
1763 | int i = 0, j, k, m; | 1769 | int i = 0, j, k, m; |
1764 | bool match = FALSE; | 1770 | bool match = FALSE; |
1765 | 1771 | ||
1766 | mmMatchedLen = -1; | 1772 | mmMatchedLen = -1; |
1767 | mmCurStack[0] = InitialState; | 1773 | mmCurStack[0] = InitialState; |
1768 | 1774 | ||
1769 | #ifndef QT_NO_REGEXP_CAPTURE | 1775 | #ifndef QT_NO_REGEXP_CAPTURE |
1770 | if ( ncap > 0 ) { | 1776 | if ( ncap > 0 ) { |
1771 | for ( j = 0; j < ncap; j++ ) { | 1777 | for ( j = 0; j < ncap; j++ ) { |
1772 | mmCurCapBegin[j] = EmptyCapture; | 1778 | mmCurCapBegin[j] = EmptyCapture; |
1773 | mmCurCapEnd[j] = EmptyCapture; | 1779 | mmCurCapEnd[j] = EmptyCapture; |
1774 | } | 1780 | } |
1775 | } | 1781 | } |
1776 | #endif | 1782 | #endif |
1777 | 1783 | ||
1778 | #ifndef QT_NO_REGEXP_BACKREF | 1784 | #ifndef QT_NO_REGEXP_BACKREF |
1779 | int *zzZ = 0; | 1785 | int *zzZ = 0; |
1780 | 1786 | ||
1781 | while ( (ncur > 0 || mmSleeping.count() > 0) && i <= mmLen - mmPos && | 1787 | while ( (ncur > 0 || mmSleeping.count() > 0) && i <= mmLen - mmPos && |
1782 | !match ) | 1788 | !match ) |
1783 | #else | 1789 | #else |
1784 | while ( ncur > 0 && i <= mmLen - mmPos && !match ) | 1790 | while ( ncur > 0 && i <= mmLen - mmPos && !match ) |
1785 | #endif | 1791 | #endif |
1786 | { | 1792 | { |
1787 | int ch = ( i < mmLen - mmPos ) ? mmIn[mmPos + i].unicode() : 0; | 1793 | int ch = ( i < mmLen - mmPos ) ? mmIn[mmPos + i].unicode() : 0; |
1788 | for ( j = 0; j < ncur; j++ ) { | 1794 | for ( j = 0; j < ncur; j++ ) { |
1789 | int cur = mmCurStack[j]; | 1795 | int cur = mmCurStack[j]; |
1790 | State *scur = s[cur]; | 1796 | State *scur = s[cur]; |
1791 | QArray<int>& outs = scur->outs; | 1797 | QArray<int>& outs = scur->outs; |
1792 | for ( k = 0; k < (int) outs.size(); k++ ) { | 1798 | for ( k = 0; k < (int) outs.size(); k++ ) { |
1793 | int next = outs[k]; | 1799 | int next = outs[k]; |
1794 | State *snext = s[next]; | 1800 | State *snext = s[next]; |
1795 | bool in = TRUE; | 1801 | bool in = TRUE; |
1796 | #ifndef QT_NO_REGEXP_BACKREF | 1802 | #ifndef QT_NO_REGEXP_BACKREF |
1797 | int needSomeSleep = 0; | 1803 | int needSomeSleep = 0; |
1798 | #endif | 1804 | #endif |
1799 | 1805 | ||
1800 | /* | 1806 | /* |
1801 | First, check if the anchors are anchored properly. | 1807 | First, check if the anchors are anchored properly. |
1802 | */ | 1808 | */ |
1803 | if ( scur->anchors != 0 ) { | 1809 | if ( scur->anchors != 0 ) { |
1804 | int a = at( *scur->anchors, next ); | 1810 | int a = at( *scur->anchors, next ); |
1805 | if ( a != 0 && !testAnchor(i, a, mmCurCapBegin + j * ncap) ) | 1811 | if ( a != 0 && !testAnchor(i, a, mmCurCapBegin + j * ncap) ) |
1806 | in = FALSE; | 1812 | in = FALSE; |
1807 | } | 1813 | } |
1808 | /* | 1814 | /* |
1809 | If indeed they are, check if the input character is correct | 1815 | If indeed they are, check if the input character is correct |
1810 | for this transition. | 1816 | for this transition. |
1811 | */ | 1817 | */ |
1812 | if ( in ) { | 1818 | if ( in ) { |
1813 | m = snext->match; | 1819 | m = snext->match; |
1814 | if ( (m & (CharClassBit | BackRefBit)) == 0 ) { | 1820 | if ( (m & (CharClassBit | BackRefBit)) == 0 ) { |
1815 | if ( cs ) | 1821 | if ( cs ) |
1816 | in = ( m == ch ); | 1822 | in = ( m == ch ); |
1817 | else | 1823 | else |
1818 | in = ( QChar(m).lower() == QChar(ch).lower() ); | 1824 | in = ( QChar(m).lower() == QChar(ch).lower() ); |
1819 | } else if ( next == FinalState ) { | 1825 | } else if ( next == FinalState ) { |
1820 | mmMatchedLen = i; | 1826 | mmMatchedLen = i; |
1821 | match = mmMinimal; | 1827 | match = mmMinimal; |
1822 | in = TRUE; | 1828 | in = TRUE; |
1823 | } else if ( (m & CharClassBit) != 0 ) { | 1829 | } else if ( (m & CharClassBit) != 0 ) { |
1824 | #ifndef QT_NO_REGEXP_CCLASS | 1830 | #ifndef QT_NO_REGEXP_CCLASS |
1825 | const CharClass *cc = cl[m ^ CharClassBit]; | 1831 | const CharClass *cc = cl[m ^ CharClassBit]; |
1826 | if ( cs ) | 1832 | if ( cs ) |
1827 | in = cc->in( ch ); | 1833 | in = cc->in( ch ); |
1828 | else if ( cc->negative() ) | 1834 | else if ( cc->negative() ) |
1829 | in = cc->in( QChar(ch).lower() ) && | 1835 | in = cc->in( QChar(ch).lower() ) && |
1830 | cc->in( QChar(ch).upper() ); | 1836 | cc->in( QChar(ch).upper() ); |
1831 | else | 1837 | else |
1832 | in = cc->in( QChar(ch).lower() ) || | 1838 | in = cc->in( QChar(ch).lower() ) || |
1833 | cc->in( QChar(ch).upper() ); | 1839 | cc->in( QChar(ch).upper() ); |
1834 | #endif | 1840 | #endif |
1835 | #ifndef QT_NO_REGEXP_BACKREF | 1841 | #ifndef QT_NO_REGEXP_BACKREF |
1836 | } else { /* ( (m & BackRefBit) != 0 ) */ | 1842 | } else { /* ( (m & BackRefBit) != 0 ) */ |
1837 | int bref = m ^ BackRefBit; | 1843 | int bref = m ^ BackRefBit; |
1838 | int ell = j * ncap + ( bref - 1 ); | 1844 | int ell = j * ncap + ( bref - 1 ); |
1839 | 1845 | ||
1840 | in = bref <= ncap && mmCurCapBegin[ell] != EmptyCapture; | 1846 | in = bref <= ncap && mmCurCapBegin[ell] != EmptyCapture; |
1841 | if ( in ) { | 1847 | if ( in ) { |
1842 | if ( cs ) | 1848 | if ( cs ) |
1843 | in = ( mmIn[mmPos + mmCurCapBegin[ell]] | 1849 | in = ( mmIn[mmPos + mmCurCapBegin[ell]] |
1844 | == QChar(ch) ); | 1850 | == QChar(ch) ); |
1845 | else | 1851 | else |
1846 | in = ( mmIn[mmPos + mmCurCapBegin[ell]].lower() | 1852 | in = ( mmIn[mmPos + mmCurCapBegin[ell]].lower() |
1847 | == QChar(ch).lower() ); | 1853 | == QChar(ch).lower() ); |
1848 | } | 1854 | } |
1849 | 1855 | ||
1850 | if ( in ) { | 1856 | if ( in ) { |
1851 | int delta; | 1857 | int delta; |
1852 | if ( mmCurCapEnd[ell] == EmptyCapture ) | 1858 | if ( mmCurCapEnd[ell] == EmptyCapture ) |
1853 | delta = i - mmCurCapBegin[ell]; | 1859 | delta = i - mmCurCapBegin[ell]; |
1854 | else | 1860 | else |
1855 | delta = mmCurCapEnd[ell] - mmCurCapBegin[ell]; | 1861 | delta = mmCurCapEnd[ell] - mmCurCapBegin[ell]; |
1856 | 1862 | ||
1857 | in = ( delta <= mmLen - mmPos ); | 1863 | in = ( delta <= mmLen - mmPos ); |
1858 | if ( in && delta > 1 ) { | 1864 | if ( in && delta > 1 ) { |
1859 | int n; | 1865 | int n; |
1860 | if ( cs ) { | 1866 | if ( cs ) { |
1861 | for ( n = 1; n < delta; n++ ) { | 1867 | for ( n = 1; n < delta; n++ ) { |
1862 | if ( mmIn[mmPos + | 1868 | if ( mmIn[mmPos + |
1863 | mmCurCapBegin[ell] + n] != | 1869 | mmCurCapBegin[ell] + n] != |
1864 | mmIn[mmPos + i + n] ) | 1870 | mmIn[mmPos + i + n] ) |
1865 | break; | 1871 | break; |
1866 | } | 1872 | } |
1867 | } else { | 1873 | } else { |
1868 | for ( n = 1; n < delta; n++ ) { | 1874 | for ( n = 1; n < delta; n++ ) { |
1869 | QChar a = mmIn[mmPos + | 1875 | QChar a = mmIn[mmPos + |
1870 | mmCurCapBegin[ell] + n]; | 1876 | mmCurCapBegin[ell] + n]; |
1871 | QChar b = mmIn[mmPos + i + n]; | 1877 | QChar b = mmIn[mmPos + i + n]; |
1872 | if ( a.lower() != b.lower() ) | 1878 | if ( a.lower() != b.lower() ) |
1873 | break; | 1879 | break; |
1874 | } | 1880 | } |
1875 | } | 1881 | } |
1876 | in = ( n == delta ); | 1882 | in = ( n == delta ); |
1877 | if ( in ) | 1883 | if ( in ) |
1878 | needSomeSleep = delta - 1; | 1884 | needSomeSleep = delta - 1; |
1879 | } | 1885 | } |
1880 | } | 1886 | } |
1881 | #endif | 1887 | #endif |
1882 | } | 1888 | } |
1883 | } | 1889 | } |
1884 | 1890 | ||
1885 | /* | 1891 | /* |
1886 | All is right. We must now update our data structures. | 1892 | All is right. We must now update our data structures. |
1887 | */ | 1893 | */ |
1888 | if ( in ) { | 1894 | if ( in ) { |
1889 | #ifndef QT_NO_REGEXP_CAPTURE | 1895 | #ifndef QT_NO_REGEXP_CAPTURE |
1890 | int *capBegin, *capEnd; | 1896 | int *capBegin, *capEnd; |
1891 | #endif | 1897 | #endif |
1892 | /* | 1898 | /* |
1893 | If the next state was not encountered yet, all is fine. | 1899 | If the next state was not encountered yet, all is fine. |
1894 | */ | 1900 | */ |
1895 | if ( (m = mmInNextStack[next]) == -1 ) { | 1901 | if ( (m = mmInNextStack[next]) == -1 ) { |
1896 | m = nnext++; | 1902 | m = nnext++; |
1897 | mmNextStack[m] = next; | 1903 | mmNextStack[m] = next; |
1898 | mmInNextStack[next] = m; | 1904 | mmInNextStack[next] = m; |
1899 | #ifndef QT_NO_REGEXP_CAPTURE | 1905 | #ifndef QT_NO_REGEXP_CAPTURE |
1900 | capBegin = mmNextCapBegin + m * ncap; | 1906 | capBegin = mmNextCapBegin + m * ncap; |
1901 | capEnd = mmNextCapEnd + m * ncap; | 1907 | capEnd = mmNextCapEnd + m * ncap; |
1902 | 1908 | ||
1903 | /* | 1909 | /* |
1904 | Otherwise, we'll first maintain captures in temporary | 1910 | Otherwise, we'll first maintain captures in temporary |
1905 | arrays, and decide at the end whether it's best to keep | 1911 | arrays, and decide at the end whether it's best to keep |
1906 | the previous capture zones or the new ones. | 1912 | the previous capture zones or the new ones. |
1907 | */ | 1913 | */ |
1908 | } else { | 1914 | } else { |
1909 | capBegin = mmTempCapBegin; | 1915 | capBegin = mmTempCapBegin; |
1910 | capEnd = mmTempCapEnd; | 1916 | capEnd = mmTempCapEnd; |
1911 | #endif | 1917 | #endif |
1912 | } | 1918 | } |
1913 | 1919 | ||
1914 | #ifndef QT_NO_REGEXP_CAPTURE | 1920 | #ifndef QT_NO_REGEXP_CAPTURE |
1915 | /* | 1921 | /* |
1916 | Updating the capture zones is much of a task. | 1922 | Updating the capture zones is much of a task. |
1917 | */ | 1923 | */ |
1918 | if ( ncap > 0 ) { | 1924 | if ( ncap > 0 ) { |
1919 | memcpy( capBegin, mmCurCapBegin + j * ncap, | 1925 | memcpy( capBegin, mmCurCapBegin + j * ncap, |
1920 | ncap * sizeof(int) ); | 1926 | ncap * sizeof(int) ); |
1921 | memcpy( capEnd, mmCurCapEnd + j * ncap, | 1927 | memcpy( capEnd, mmCurCapEnd + j * ncap, |
1922 | ncap * sizeof(int) ); | 1928 | ncap * sizeof(int) ); |
1923 | int c = scur->atom, n = snext->atom; | 1929 | int c = scur->atom, n = snext->atom; |
1924 | int p = -1, q = -1; | 1930 | int p = -1, q = -1; |
1925 | int cap; | 1931 | int cap; |
1926 | 1932 | ||
1927 | /* | 1933 | /* |
1928 | Lemma 1. For any x in the range [0..nf), we have | 1934 | Lemma 1. For any x in the range [0..nf), we have |
1929 | f[x].parent < x. | 1935 | f[x].parent < x. |
1930 | 1936 | ||
1931 | Proof. By looking at startAtom(), it is clear that | 1937 | Proof. By looking at startAtom(), it is clear that |
1932 | cf < nf holds all the time, and thus that | 1938 | cf < nf holds all the time, and thus that |
1933 | f[nf].parent < nf. | 1939 | f[nf].parent < nf. |
1934 | */ | 1940 | */ |
1935 | 1941 | ||
1936 | /* | 1942 | /* |
1937 | If we are reentering an atom, we empty all capture | 1943 | If we are reentering an atom, we empty all capture |
1938 | zones inside it. | 1944 | zones inside it. |
1939 | */ | 1945 | */ |
1940 | if ( scur->reenter != 0 && | 1946 | if ( scur->reenter != 0 && |
1941 | (q = at(*scur->reenter, next)) != 0 ) { | 1947 | (q = at(*scur->reenter, next)) != 0 ) { |
1942 | QBitArray b; | 1948 | QBitArray b; |
1943 | b.fill( FALSE, nf ); | 1949 | b.fill( FALSE, nf ); |
1944 | b.setBit( q, TRUE ); | 1950 | b.setBit( q, TRUE ); |
1945 | for ( int ell = q + 1; ell < nf; ell++ ) { | 1951 | for ( int ell = q + 1; ell < nf; ell++ ) { |
1946 | if ( b.testBit(f[ell].parent) ) { | 1952 | if ( b.testBit(f[ell].parent) ) { |
1947 | b.setBit( ell, TRUE ); | 1953 | b.setBit( ell, TRUE ); |
1948 | cap = f[ell].capture; | 1954 | cap = f[ell].capture; |
1949 | if ( cap >= 0 ) { | 1955 | if ( cap >= 0 ) { |
1950 | capBegin[cap] = EmptyCapture; | 1956 | capBegin[cap] = EmptyCapture; |
1951 | capEnd[cap] = EmptyCapture; | 1957 | capEnd[cap] = EmptyCapture; |
1952 | } | 1958 | } |
1953 | } | 1959 | } |
1954 | } | 1960 | } |
1955 | p = f[q].parent; | 1961 | p = f[q].parent; |
1956 | 1962 | ||
1957 | /* | 1963 | /* |
1958 | Otherwise, close the capture zones we are leaving. | 1964 | Otherwise, close the capture zones we are leaving. |
1959 | We are leaving f[c].capture, f[f[c].parent].capture, | 1965 | We are leaving f[c].capture, f[f[c].parent].capture, |
1960 | f[f[f[c].parent].parent].capture, ..., until | 1966 | f[f[f[c].parent].parent].capture, ..., until |
1961 | f[x].capture, with x such that f[x].parent is the | 1967 | f[x].capture, with x such that f[x].parent is the |
1962 | youngest common ancestor for c and n. | 1968 | youngest common ancestor for c and n. |
1963 | 1969 | ||
1964 | We go up along c's and n's ancestry until we find x. | 1970 | We go up along c's and n's ancestry until we find x. |
1965 | */ | 1971 | */ |
1966 | } else { | 1972 | } else { |
1967 | p = c; | 1973 | p = c; |
1968 | q = n; | 1974 | q = n; |
1969 | while ( p != q ) { | 1975 | while ( p != q ) { |
1970 | if ( p > q ) { | 1976 | if ( p > q ) { |
1971 | cap = f[p].capture; | 1977 | cap = f[p].capture; |
1972 | if ( cap >= 0 ) { | 1978 | if ( cap >= 0 ) { |
1973 | if ( capBegin[cap] == i ) { | 1979 | if ( capBegin[cap] == i ) { |
1974 | capBegin[cap] = EmptyCapture; | 1980 | capBegin[cap] = EmptyCapture; |
1975 | capEnd[cap] = EmptyCapture; | 1981 | capEnd[cap] = EmptyCapture; |
1976 | } else { | 1982 | } else { |
1977 | capEnd[cap] = i; | 1983 | capEnd[cap] = i; |
1978 | } | 1984 | } |
1979 | } | 1985 | } |
1980 | p = f[p].parent; | 1986 | p = f[p].parent; |
1981 | } else { | 1987 | } else { |
1982 | q = f[q].parent; | 1988 | q = f[q].parent; |
1983 | } | 1989 | } |
1984 | } | 1990 | } |
1985 | } | 1991 | } |
1986 | 1992 | ||
1987 | /* | 1993 | /* |
1988 | In any case, we now open the capture zones we are | 1994 | In any case, we now open the capture zones we are |
1989 | entering. We work upwards from n until we reach p | 1995 | entering. We work upwards from n until we reach p |
1990 | (the parent of the atom we reenter or the youngest | 1996 | (the parent of the atom we reenter or the youngest |
1991 | common ancestor). | 1997 | common ancestor). |
1992 | */ | 1998 | */ |
1993 | while ( n > p ) { | 1999 | while ( n > p ) { |
1994 | cap = f[n].capture; | 2000 | cap = f[n].capture; |
1995 | if ( cap >= 0 ) { | 2001 | if ( cap >= 0 ) { |
1996 | capBegin[cap] = i; | 2002 | capBegin[cap] = i; |
1997 | capEnd[cap] = EmptyCapture; | 2003 | capEnd[cap] = EmptyCapture; |
1998 | } | 2004 | } |
1999 | n = f[n].parent; | 2005 | n = f[n].parent; |
2000 | } | 2006 | } |
2001 | /* | 2007 | /* |
2002 | If the next state was already in mmNextStack, we must | 2008 | If the next state was already in mmNextStack, we must |
2003 | choose carefully which capture zones we want to keep. | 2009 | choose carefully which capture zones we want to keep. |
2004 | */ | 2010 | */ |
2005 | if ( capBegin == mmTempCapBegin && | 2011 | if ( capBegin == mmTempCapBegin && |
2006 | isBetterCapture(capBegin, capEnd, | 2012 | isBetterCapture(capBegin, capEnd, |
2007 | mmNextCapBegin + m * ncap, | 2013 | mmNextCapBegin + m * ncap, |
2008 | mmNextCapEnd + m * ncap) ) { | 2014 | mmNextCapEnd + m * ncap) ) { |
2009 | memcpy( mmNextCapBegin + m * ncap, capBegin, | 2015 | memcpy( mmNextCapBegin + m * ncap, capBegin, |
2010 | ncap * sizeof(int) ); | 2016 | ncap * sizeof(int) ); |
2011 | memcpy( mmNextCapEnd + m * ncap, capEnd, | 2017 | memcpy( mmNextCapEnd + m * ncap, capEnd, |
2012 | ncap * sizeof(int) ); | 2018 | ncap * sizeof(int) ); |
2013 | } | 2019 | } |
2014 | } | 2020 | } |
2015 | #ifndef QT_NO_REGEXP_BACKREF | 2021 | #ifndef QT_NO_REGEXP_BACKREF |
2016 | /* | 2022 | /* |
2017 | We are done with updating the capture zones. It's now | 2023 | We are done with updating the capture zones. It's now |
2018 | time to put the next state to sleep, if it needs to, and | 2024 | time to put the next state to sleep, if it needs to, and |
2019 | to remove it from mmNextStack. | 2025 | to remove it from mmNextStack. |
2020 | */ | 2026 | */ |
2021 | if ( needSomeSleep > 0 ) { | 2027 | if ( needSomeSleep > 0 ) { |
2022 | zzZ = new int[1 + 2 * ncap]; | 2028 | zzZ = new int[1 + 2 * ncap]; |
2023 | zzZ[0] = next; | 2029 | zzZ[0] = next; |
2024 | if ( ncap > 0 ) { | 2030 | if ( ncap > 0 ) { |
2025 | memcpy( zzZ + 1, capBegin, ncap * sizeof(int) ); | 2031 | memcpy( zzZ + 1, capBegin, ncap * sizeof(int) ); |
2026 | memcpy( zzZ + 1 + ncap, capEnd, | 2032 | memcpy( zzZ + 1 + ncap, capEnd, |
2027 | ncap * sizeof(int) ); | 2033 | ncap * sizeof(int) ); |
2028 | } | 2034 | } |
2029 | mmInNextStack[mmNextStack[--nnext]] = -1; | 2035 | mmInNextStack[mmNextStack[--nnext]] = -1; |
2030 | mmSleeping.insert( i + needSomeSleep, zzZ ); | 2036 | mmSleeping.insert( i + needSomeSleep, zzZ ); |
2031 | } | 2037 | } |
2032 | #endif | 2038 | #endif |
2033 | #endif | 2039 | #endif |
2034 | } | 2040 | } |
2035 | } | 2041 | } |
2036 | } | 2042 | } |
2037 | #ifndef QT_NO_REGEXP_CAPTURE | 2043 | #ifndef QT_NO_REGEXP_CAPTURE |
2038 | /* | 2044 | /* |
2039 | If we reached the final state, hurray! Copy the captured zone. | 2045 | If we reached the final state, hurray! Copy the captured zone. |
2040 | */ | 2046 | */ |
2041 | if ( ncap > 0 && (m = mmInNextStack[FinalState]) != -1 ) { | 2047 | if ( ncap > 0 && (m = mmInNextStack[FinalState]) != -1 ) { |
2042 | memcpy( mmCapBegin, mmNextCapBegin + m * ncap, ncap * sizeof(int) ); | 2048 | memcpy( mmCapBegin, mmNextCapBegin + m * ncap, ncap * sizeof(int) ); |
2043 | memcpy( mmCapEnd, mmNextCapEnd + m * ncap, ncap * sizeof(int) ); | 2049 | memcpy( mmCapEnd, mmNextCapEnd + m * ncap, ncap * sizeof(int) ); |
2044 | } | 2050 | } |
2045 | #ifndef QT_NO_REGEXP_BACKREF | 2051 | #ifndef QT_NO_REGEXP_BACKREF |
2046 | /* | 2052 | /* |
2047 | It's time to wake up the sleepers. | 2053 | It's time to wake up the sleepers. |
2048 | */ | 2054 | */ |
2049 | if ( mmSleeping.count() > 0 ) { | 2055 | if ( mmSleeping.count() > 0 ) { |
2050 | while ( (zzZ = mmSleeping.take(i)) != 0 ) { | 2056 | while ( (zzZ = mmSleeping.take(i)) != 0 ) { |
2051 | int next = zzZ[0]; | 2057 | int next = zzZ[0]; |
2052 | int *capBegin = zzZ + 1; | 2058 | int *capBegin = zzZ + 1; |
2053 | int *capEnd = zzZ + 1 + ncap; | 2059 | int *capEnd = zzZ + 1 + ncap; |
2054 | bool copyOver = TRUE; | 2060 | bool copyOver = TRUE; |
2055 | 2061 | ||
2056 | if ( (m = mmInNextStack[zzZ[0]]) == -1 ) { | 2062 | if ( (m = mmInNextStack[zzZ[0]]) == -1 ) { |
2057 | m = nnext++; | 2063 | m = nnext++; |
2058 | mmNextStack[m] = next; | 2064 | mmNextStack[m] = next; |
2059 | mmInNextStack[next] = m; | 2065 | mmInNextStack[next] = m; |
2060 | } else { | 2066 | } else { |
2061 | copyOver = isBetterCapture( mmNextCapBegin + m * ncap, | 2067 | copyOver = isBetterCapture( mmNextCapBegin + m * ncap, |
2062 | mmNextCapEnd + m * ncap, | 2068 | mmNextCapEnd + m * ncap, |
2063 | capBegin, capEnd ); | 2069 | capBegin, capEnd ); |
2064 | } | 2070 | } |
2065 | if ( copyOver ) { | 2071 | if ( copyOver ) { |
2066 | memcpy( mmNextCapBegin + m * ncap, capBegin, | 2072 | memcpy( mmNextCapBegin + m * ncap, capBegin, |
2067 | ncap * sizeof(int) ); | 2073 | ncap * sizeof(int) ); |
2068 | memcpy( mmNextCapEnd + m * ncap, capEnd, | 2074 | memcpy( mmNextCapEnd + m * ncap, capEnd, |
2069 | ncap * sizeof(int) ); | 2075 | ncap * sizeof(int) ); |
2070 | } | 2076 | } |
2071 | delete[] zzZ; | 2077 | delete[] zzZ; |
2072 | } | 2078 | } |
2073 | } | 2079 | } |
2074 | #endif | 2080 | #endif |
2075 | #endif | 2081 | #endif |
2076 | for ( j = 0; j < nnext; j++ ) | 2082 | for ( j = 0; j < nnext; j++ ) |
2077 | mmInNextStack[mmNextStack[j]] = -1; | 2083 | mmInNextStack[mmNextStack[j]] = -1; |
2078 | 2084 | ||
2079 | qSwap( mmCurStack, mmNextStack ); | 2085 | qSwap( mmCurStack, mmNextStack ); |
2080 | #ifndef QT_NO_REGEXP_CAPTURE | 2086 | #ifndef QT_NO_REGEXP_CAPTURE |
2081 | qSwap( mmCurCapBegin, mmNextCapBegin ); | 2087 | qSwap( mmCurCapBegin, mmNextCapBegin ); |
2082 | qSwap( mmCurCapEnd, mmNextCapEnd ); | 2088 | qSwap( mmCurCapEnd, mmNextCapEnd ); |
2083 | #endif | 2089 | #endif |
2084 | ncur = nnext; | 2090 | ncur = nnext; |
2085 | nnext = 0; | 2091 | nnext = 0; |
2086 | i++; | 2092 | i++; |
2087 | } | 2093 | } |
2088 | 2094 | ||
2089 | #ifndef QT_NO_REGEXP_BACKREF | 2095 | #ifndef QT_NO_REGEXP_BACKREF |
2090 | /* | 2096 | /* |
2091 | If minimal matching is enabled, we might have some sleepers left. | 2097 | If minimal matching is enabled, we might have some sleepers left. |
2092 | */ | 2098 | */ |
2093 | while ( !mmSleeping.isEmpty() ) { | 2099 | while ( !mmSleeping.isEmpty() ) { |
2094 | zzZ = mmSleeping.take( *QIntDictIterator<int>(mmSleeping) ); | 2100 | zzZ = mmSleeping.take( *QIntDictIterator<int>(mmSleeping) ); |
2095 | delete[] zzZ; | 2101 | delete[] zzZ; |
2096 | } | 2102 | } |
2097 | #endif | 2103 | #endif |
2098 | 2104 | ||
2099 | match = ( mmMatchedLen >= 0 ); | 2105 | match = ( mmMatchedLen >= 0 ); |
2100 | if ( !match ) | 2106 | if ( !match ) |
2101 | mmMatchedLen = i - 1; | 2107 | mmMatchedLen = i - 1; |
2102 | return match; | 2108 | return match; |
2103 | } | 2109 | } |
2104 | 2110 | ||
2105 | #ifndef QT_NO_REGEXP_CCLASS | 2111 | #ifndef QT_NO_REGEXP_CCLASS |
2106 | 2112 | ||
2107 | QRegExpEngine::CharClass::CharClass() | 2113 | QRegExpEngine::CharClass::CharClass() |
2108 | : c( 0 ), n( FALSE ) | 2114 | : c( 0 ), n( FALSE ) |
2109 | #ifndef QT_NO_REGEXP_OPTIM | 2115 | #ifndef QT_NO_REGEXP_OPTIM |
2110 | , occ1( *noOccurrences ) | 2116 | , occ1( *noOccurrences ) |
2111 | #endif | 2117 | #endif |
2112 | { | 2118 | { |
2113 | } | 2119 | } |
2114 | 2120 | ||
2115 | QRegExpEngine::CharClass& QRegExpEngine::CharClass::operator=( | 2121 | QRegExpEngine::CharClass& QRegExpEngine::CharClass::operator=( |
2116 | const CharClass& cc ) | 2122 | const CharClass& cc ) |
2117 | { | 2123 | { |
2118 | c = cc.c; | 2124 | c = cc.c; |
2119 | r = cc.r.copy(); | 2125 | r = cc.r.copy(); |
2120 | n = cc.n; | 2126 | n = cc.n; |
2121 | #ifndef QT_NO_REGEXP_OPTIM | 2127 | #ifndef QT_NO_REGEXP_OPTIM |
2122 | occ1 = cc.occ1; | 2128 | occ1 = cc.occ1; |
2123 | #endif | 2129 | #endif |
2124 | return *this; | 2130 | return *this; |
2125 | } | 2131 | } |
2126 | 2132 | ||
2127 | void QRegExpEngine::CharClass::clear() | 2133 | void QRegExpEngine::CharClass::clear() |
2128 | { | 2134 | { |
2129 | c = 0; | 2135 | c = 0; |
2130 | r.resize( 0 ); | 2136 | r.resize( 0 ); |
2131 | n = FALSE; | 2137 | n = FALSE; |
2132 | } | 2138 | } |
2133 | 2139 | ||
2134 | void QRegExpEngine::CharClass::setNegative( bool negative ) | 2140 | void QRegExpEngine::CharClass::setNegative( bool negative ) |
2135 | { | 2141 | { |
2136 | n = negative; | 2142 | n = negative; |
2137 | #ifndef QT_NO_REGEXP_OPTIM | 2143 | #ifndef QT_NO_REGEXP_OPTIM |
2138 | occ1 = *firstOccurrenceAtZero; | 2144 | occ1 = *firstOccurrenceAtZero; |
2139 | #endif | 2145 | #endif |
2140 | } | 2146 | } |
2141 | 2147 | ||
2142 | void QRegExpEngine::CharClass::addCategories( int cats ) | 2148 | void QRegExpEngine::CharClass::addCategories( int cats ) |
2143 | { | 2149 | { |
2144 | c |= cats; | 2150 | c |= cats; |
2145 | #ifndef QT_NO_REGEXP_OPTIM | 2151 | #ifndef QT_NO_REGEXP_OPTIM |
2146 | occ1 = *firstOccurrenceAtZero; | 2152 | occ1 = *firstOccurrenceAtZero; |
2147 | #endif | 2153 | #endif |
2148 | } | 2154 | } |
2149 | 2155 | ||
2150 | void QRegExpEngine::CharClass::addRange( ushort from, ushort to ) | 2156 | void QRegExpEngine::CharClass::addRange( ushort from, ushort to ) |
2151 | { | 2157 | { |
2152 | if ( from > to ) | 2158 | if ( from > to ) |
2153 | qSwap( from, to ); | 2159 | qSwap( from, to ); |
2154 | int n = r.size(); | 2160 | int n = r.size(); |
2155 | r.resize( n + 1 ); | 2161 | r.resize( n + 1 ); |
2156 | r[n].from = from; | 2162 | r[n].from = from; |
2157 | r[n].to = to; | 2163 | r[n].to = to; |
2158 | 2164 | ||
2159 | #ifndef QT_NO_REGEXP_OPTIM | 2165 | #ifndef QT_NO_REGEXP_OPTIM |
2160 | int i; | 2166 | int i; |
2161 | 2167 | ||
2162 | if ( to - from < NumBadChars ) { | 2168 | if ( to - from < NumBadChars ) { |
2163 | occ1.detach(); | 2169 | occ1.detach(); |
2164 | if ( from % NumBadChars <= to % NumBadChars ) { | 2170 | if ( from % NumBadChars <= to % NumBadChars ) { |
2165 | for ( i = from % NumBadChars; i <= to % NumBadChars; i++ ) | 2171 | for ( i = from % NumBadChars; i <= to % NumBadChars; i++ ) |
2166 | occ1[i] = 0; | 2172 | occ1[i] = 0; |
2167 | } else { | ||
2168 | for ( i = 0; i <= to % NumBadChars; i++ ) | ||
2169 | occ1[i] = 0; | ||
2170 | for ( i = from % NumBadChars; i < NumBadChars; i++ ) | ||
2171 | occ1[i] = 0; | ||
2172 | } | ||
2173 | } else { | 2173 | } else { |
2174 | occ1 = *firstOccurrenceAtZero; | 2174 | for ( i = 0; i <= to % NumBadChars; i++ ) |
2175 | occ1[i] = 0; | ||
2176 | for ( i = from % NumBadChars; i < NumBadChars; i++ ) | ||
2177 | occ1[i] = 0; | ||
2178 | } | ||
2179 | } else { | ||
2180 | occ1 = *firstOccurrenceAtZero; | ||
2175 | } | 2181 | } |
2176 | #endif | 2182 | #endif |
2177 | } | 2183 | } |
2178 | 2184 | ||
2179 | bool QRegExpEngine::CharClass::in( QChar ch ) const | 2185 | bool QRegExpEngine::CharClass::in( QChar ch ) const |
2180 | { | 2186 | { |
2181 | #ifndef QT_NO_REGEXP_OPTIM | 2187 | #ifndef QT_NO_REGEXP_OPTIM |
2182 | if ( occ1[BadChar(ch)] == NoOccurrence ) | 2188 | if ( occ1[BadChar(ch)] == NoOccurrence ) |
2183 | return n; | 2189 | return n; |
2184 | #endif | 2190 | #endif |
2185 | 2191 | ||
2186 | if ( c != 0 && (c & (1 << (int) ch.category())) != 0 ) | 2192 | if ( c != 0 && (c & (1 << (int) ch.category())) != 0 ) |
2187 | return !n; | 2193 | return !n; |
2188 | for ( int i = 0; i < (int) r.size(); i++ ) { | 2194 | for ( int i = 0; i < (int) r.size(); i++ ) { |
2189 | if ( ch.unicode() >= r[i].from && ch.unicode() <= r[i].to ) | 2195 | if ( ch.unicode() >= r[i].from && ch.unicode() <= r[i].to ) |
2190 | return !n; | 2196 | return !n; |
2191 | } | 2197 | } |
2192 | return n; | 2198 | return n; |
2193 | } | 2199 | } |
2194 | 2200 | ||
2195 | #if defined(QT_DEBUG) | 2201 | #if defined(QT_DEBUG) |
2196 | void QRegExpEngine::CharClass::dump() const | 2202 | void QRegExpEngine::CharClass::dump() const |
2197 | { | 2203 | { |
2198 | int i; | 2204 | int i; |
2199 | qDebug( " %stive character class", n ? "nega" : "posi" ); | 2205 | odebug << " " << (n ? "nega" : "posi") << "tive character class" << oendl; |
2200 | #ifndef QT_NO_REGEXP_CCLASS | 2206 | #ifndef QT_NO_REGEXP_CCLASS |
2201 | if ( c != 0 ) | 2207 | if ( c != 0 ) |
2202 | qDebug( " categories 0x%.8x", c ); | 2208 | odebug << QString().sprintf(" categories 0x%.8x", c ) << oendl; |
2203 | #endif | 2209 | #endif |
2204 | for ( i = 0; i < (int) r.size(); i++ ) | 2210 | for ( i = 0; i < (int) r.size(); i++ ) |
2205 | qDebug( " 0x%.4x through 0x%.4x", r[i].from, r[i].to ); | 2211 | odebug << QString().sprintf(" 0x%.4x through 0x%.4x", r[i].from, r[i].to ) << oendl; |
2206 | } | 2212 | } |
2207 | #endif | 2213 | #endif |
2208 | #endif | 2214 | #endif |
2209 | 2215 | ||
2210 | QRegExpEngine::Box::Box( QRegExpEngine *engine ) | 2216 | QRegExpEngine::Box::Box( QRegExpEngine *engine ) |
2211 | : eng( engine ), skipanchors( 0 ) | 2217 | : eng( engine ), skipanchors( 0 ) |
2212 | #ifndef QT_NO_REGEXP_OPTIM | 2218 | #ifndef QT_NO_REGEXP_OPTIM |
2213 | , earlyStart( 0 ), lateStart( 0 ), maxl( 0 ), occ1( *noOccurrences ) | 2219 | , earlyStart( 0 ), lateStart( 0 ), maxl( 0 ), occ1( *noOccurrences ) |
2214 | #endif | 2220 | #endif |
2215 | { | 2221 | { |
2216 | minl = 0; | 2222 | minl = 0; |
2217 | } | 2223 | } |
2218 | 2224 | ||
2219 | QRegExpEngine::Box& QRegExpEngine::Box::operator=( const Box& b ) | 2225 | QRegExpEngine::Box& QRegExpEngine::Box::operator=( const Box& b ) |
2220 | { | 2226 | { |
2221 | eng = b.eng; | 2227 | eng = b.eng; |
2222 | ls = b.ls; | 2228 | ls = b.ls; |
2223 | rs = b.rs; | 2229 | rs = b.rs; |
2224 | lanchors = b.lanchors; | 2230 | lanchors = b.lanchors; |
2225 | ranchors = b.ranchors; | 2231 | ranchors = b.ranchors; |
2226 | skipanchors = b.skipanchors; | 2232 | skipanchors = b.skipanchors; |
2227 | #ifndef QT_NO_REGEXP_OPTIM | 2233 | #ifndef QT_NO_REGEXP_OPTIM |
2228 | earlyStart = b.earlyStart; | 2234 | earlyStart = b.earlyStart; |
2229 | lateStart = b.lateStart; | 2235 | lateStart = b.lateStart; |
2230 | str = b.str; | 2236 | str = b.str; |
2231 | leftStr = b.leftStr; | 2237 | leftStr = b.leftStr; |
2232 | rightStr = b.rightStr; | 2238 | rightStr = b.rightStr; |
2233 | maxl = b.maxl; | 2239 | maxl = b.maxl; |
2234 | occ1 = b.occ1; | 2240 | occ1 = b.occ1; |
2235 | #endif | 2241 | #endif |
2236 | minl = b.minl; | 2242 | minl = b.minl; |
2237 | return *this; | 2243 | return *this; |
2238 | } | 2244 | } |
2239 | 2245 | ||
2240 | void QRegExpEngine::Box::set( QChar ch ) | 2246 | void QRegExpEngine::Box::set( QChar ch ) |
2241 | { | 2247 | { |
2242 | ls.resize( 1 ); | 2248 | ls.resize( 1 ); |
2243 | ls[0] = eng->createState( ch ); | 2249 | ls[0] = eng->createState( ch ); |
2244 | rs = ls; | 2250 | rs = ls; |
2245 | rs.detach(); | 2251 | rs.detach(); |
2246 | #ifndef QT_NO_REGEXP_OPTIM | 2252 | #ifndef QT_NO_REGEXP_OPTIM |
2247 | str = ch; | 2253 | str = ch; |
2248 | leftStr = ch; | 2254 | leftStr = ch; |
2249 | rightStr = ch; | 2255 | rightStr = ch; |
2250 | maxl = 1; | 2256 | maxl = 1; |
2251 | occ1.detach(); | 2257 | occ1.detach(); |
2252 | occ1[BadChar(ch)] = 0; | 2258 | occ1[BadChar(ch)] = 0; |
2253 | #endif | 2259 | #endif |
2254 | minl = 1; | 2260 | minl = 1; |
2255 | } | 2261 | } |
2256 | 2262 | ||
2257 | void QRegExpEngine::Box::set( const CharClass& cc ) | 2263 | void QRegExpEngine::Box::set( const CharClass& cc ) |
2258 | { | 2264 | { |
2259 | ls.resize( 1 ); | 2265 | ls.resize( 1 ); |
2260 | ls[0] = eng->createState( cc ); | 2266 | ls[0] = eng->createState( cc ); |
2261 | rs = ls; | 2267 | rs = ls; |
2262 | rs.detach(); | 2268 | rs.detach(); |
2263 | #ifndef QT_NO_REGEXP_OPTIM | 2269 | #ifndef QT_NO_REGEXP_OPTIM |
2264 | maxl = 1; | 2270 | maxl = 1; |
2265 | occ1 = cc.firstOccurrence(); | 2271 | occ1 = cc.firstOccurrence(); |
2266 | #endif | 2272 | #endif |
2267 | minl = 1; | 2273 | minl = 1; |
2268 | } | 2274 | } |
2269 | 2275 | ||
2270 | #ifndef QT_NO_REGEXP_BACKREF | 2276 | #ifndef QT_NO_REGEXP_BACKREF |
2271 | void QRegExpEngine::Box::set( int bref ) | 2277 | void QRegExpEngine::Box::set( int bref ) |
2272 | { | 2278 | { |
2273 | ls.resize( 1 ); | 2279 | ls.resize( 1 ); |
2274 | ls[0] = eng->createState( bref ); | 2280 | ls[0] = eng->createState( bref ); |
2275 | rs = ls; | 2281 | rs = ls; |
2276 | rs.detach(); | 2282 | rs.detach(); |
2277 | if ( bref >= 1 && bref <= MaxBackRefs ) | 2283 | if ( bref >= 1 && bref <= MaxBackRefs ) |
2278 | skipanchors = Anchor_BackRef0Empty << bref; | 2284 | skipanchors = Anchor_BackRef0Empty << bref; |
2279 | #ifndef QT_NO_REGEXP_OPTIM | 2285 | #ifndef QT_NO_REGEXP_OPTIM |
2280 | maxl = InftyLen; | 2286 | maxl = InftyLen; |
2281 | #endif | 2287 | #endif |
2282 | minl = 0; | 2288 | minl = 0; |
2283 | } | 2289 | } |
2284 | #endif | 2290 | #endif |
2285 | 2291 | ||
2286 | void QRegExpEngine::Box::cat( const Box& b ) | 2292 | void QRegExpEngine::Box::cat( const Box& b ) |
2287 | { | 2293 | { |
2288 | eng->addCatTransitions( rs, b.ls ); | 2294 | eng->addCatTransitions( rs, b.ls ); |
2289 | addAnchorsToEngine( b ); | 2295 | addAnchorsToEngine( b ); |
2290 | if ( minl == 0 ) { | 2296 | if ( minl == 0 ) { |
2291 | mergeInto( &lanchors, b.lanchors ); | 2297 | mergeInto( &lanchors, b.lanchors ); |
2292 | if ( skipanchors != 0 ) { | 2298 | if ( skipanchors != 0 ) { |
2293 | for ( int i = 0; i < (int) b.ls.size(); i++ ) { | 2299 | for ( int i = 0; i < (int) b.ls.size(); i++ ) { |
2294 | int a = eng->anchorConcatenation( at(lanchors, b.ls[i]), | 2300 | int a = eng->anchorConcatenation( at(lanchors, b.ls[i]), |
2295 | skipanchors ); | 2301 | skipanchors ); |
2296 | lanchors.insert( b.ls[i], a ); | 2302 | lanchors.insert( b.ls[i], a ); |
2297 | } | 2303 | } |
2298 | } | 2304 | } |
2299 | mergeInto( &ls, b.ls ); | 2305 | mergeInto( &ls, b.ls ); |
2300 | } | 2306 | } |
2301 | if ( b.minl == 0 ) { | 2307 | if ( b.minl == 0 ) { |
2302 | mergeInto( &ranchors, b.ranchors ); | 2308 | mergeInto( &ranchors, b.ranchors ); |
2303 | if ( b.skipanchors != 0 ) { | 2309 | if ( b.skipanchors != 0 ) { |
2304 | for ( int i = 0; i < (int) rs.size(); i++ ) { | 2310 | for ( int i = 0; i < (int) rs.size(); i++ ) { |
2305 | int a = eng->anchorConcatenation( at(ranchors, rs[i]), | 2311 | int a = eng->anchorConcatenation( at(ranchors, rs[i]), |
2306 | b.skipanchors ); | 2312 | b.skipanchors ); |
2307 | ranchors.insert( rs[i], a ); | 2313 | ranchors.insert( rs[i], a ); |
2308 | } | 2314 | } |
2309 | } | 2315 | } |
2310 | mergeInto( &rs, b.rs ); | 2316 | mergeInto( &rs, b.rs ); |
2311 | } else { | 2317 | } else { |
2312 | ranchors = b.ranchors; | 2318 | ranchors = b.ranchors; |
2313 | rs = b.rs; | 2319 | rs = b.rs; |
2314 | } | 2320 | } |
2315 | 2321 | ||
2316 | #ifndef QT_NO_REGEXP_OPTIM | 2322 | #ifndef QT_NO_REGEXP_OPTIM |
2317 | if ( maxl != InftyLen ) { | 2323 | if ( maxl != InftyLen ) { |
2318 | if ( rightStr.length() + b.leftStr.length() > | 2324 | if ( rightStr.length() + b.leftStr.length() > |
2319 | QMAX(str.length(), b.str.length()) ) { | 2325 | QMAX(str.length(), b.str.length()) ) { |
2320 | earlyStart = minl - rightStr.length(); | 2326 | earlyStart = minl - rightStr.length(); |
2321 | lateStart = maxl - rightStr.length(); | 2327 | lateStart = maxl - rightStr.length(); |
2322 | str = rightStr + b.leftStr; | 2328 | str = rightStr + b.leftStr; |
2323 | } else if ( b.str.length() > str.length() ) { | 2329 | } else if ( b.str.length() > str.length() ) { |
2324 | earlyStart = minl + b.earlyStart; | 2330 | earlyStart = minl + b.earlyStart; |
2325 | lateStart = maxl + b.lateStart; | 2331 | lateStart = maxl + b.lateStart; |
2326 | str = b.str; | 2332 | str = b.str; |
2327 | } | 2333 | } |
2328 | } | 2334 | } |
2329 | 2335 | ||
2330 | if ( (int) leftStr.length() == maxl ) | 2336 | if ( (int) leftStr.length() == maxl ) |
2331 | leftStr += b.leftStr; | 2337 | leftStr += b.leftStr; |
2332 | if ( (int) b.rightStr.length() == b.maxl ) | 2338 | if ( (int) b.rightStr.length() == b.maxl ) |
2333 | rightStr += b.rightStr; | 2339 | rightStr += b.rightStr; |
2334 | else | 2340 | else |
2335 | rightStr = b.rightStr; | 2341 | rightStr = b.rightStr; |
2336 | 2342 | ||
2337 | if ( maxl == InftyLen || b.maxl == InftyLen ) | 2343 | if ( maxl == InftyLen || b.maxl == InftyLen ) |
2338 | maxl = InftyLen; | 2344 | maxl = InftyLen; |
2339 | else | 2345 | else |
2340 | maxl += b.maxl; | 2346 | maxl += b.maxl; |
2341 | 2347 | ||
2342 | occ1.detach(); | 2348 | occ1.detach(); |
2343 | for ( int i = 0; i < NumBadChars; i++ ) { | 2349 | for ( int i = 0; i < NumBadChars; i++ ) { |
2344 | if ( b.occ1[i] != NoOccurrence && minl + b.occ1[i] < occ1[i] ) | 2350 | if ( b.occ1[i] != NoOccurrence && minl + b.occ1[i] < occ1[i] ) |
2345 | occ1[i] = minl + b.occ1[i]; | 2351 | occ1[i] = minl + b.occ1[i]; |
2346 | } | 2352 | } |
2347 | #endif | 2353 | #endif |
2348 | 2354 | ||
2349 | minl += b.minl; | 2355 | minl += b.minl; |
2350 | if ( minl == 0 ) | 2356 | if ( minl == 0 ) |
2351 | skipanchors = eng->anchorConcatenation( skipanchors, b.skipanchors ); | 2357 | skipanchors = eng->anchorConcatenation( skipanchors, b.skipanchors ); |
2352 | else | 2358 | else |
2353 | skipanchors = 0; | 2359 | skipanchors = 0; |
2354 | } | 2360 | } |
2355 | 2361 | ||
2356 | void QRegExpEngine::Box::orx( const Box& b ) | 2362 | void QRegExpEngine::Box::orx( const Box& b ) |
2357 | { | 2363 | { |
2358 | mergeInto( &ls, b.ls ); | 2364 | mergeInto( &ls, b.ls ); |
2359 | mergeInto( &lanchors, b.lanchors ); | 2365 | mergeInto( &lanchors, b.lanchors ); |
2360 | mergeInto( &rs, b.rs ); | 2366 | mergeInto( &rs, b.rs ); |
2361 | mergeInto( &ranchors, b.ranchors ); | 2367 | mergeInto( &ranchors, b.ranchors ); |
2362 | skipanchors = eng->anchorAlternation( skipanchors, b.skipanchors ); | 2368 | skipanchors = eng->anchorAlternation( skipanchors, b.skipanchors ); |
2363 | 2369 | ||
2364 | #ifndef QT_NO_REGEXP_OPTIM | 2370 | #ifndef QT_NO_REGEXP_OPTIM |
2365 | occ1.detach(); | 2371 | occ1.detach(); |
2366 | for ( int i = 0; i < NumBadChars; i++ ) { | 2372 | for ( int i = 0; i < NumBadChars; i++ ) { |
2367 | if ( occ1[i] > b.occ1[i] ) | 2373 | if ( occ1[i] > b.occ1[i] ) |
2368 | occ1[i] = b.occ1[i]; | 2374 | occ1[i] = b.occ1[i]; |
2369 | } | 2375 | } |
2370 | earlyStart = 0; | 2376 | earlyStart = 0; |
2371 | lateStart = 0; | 2377 | lateStart = 0; |
2372 | str = QString::null; | 2378 | str = QString::null; |
2373 | leftStr = QString::null; | 2379 | leftStr = QString::null; |
2374 | rightStr = QString::null; | 2380 | rightStr = QString::null; |
2375 | if ( b.maxl > maxl ) | 2381 | if ( b.maxl > maxl ) |
2376 | maxl = b.maxl; | 2382 | maxl = b.maxl; |
2377 | #endif | 2383 | #endif |
2378 | if ( b.minl < minl ) | 2384 | if ( b.minl < minl ) |
2379 | minl = b.minl; | 2385 | minl = b.minl; |
2380 | } | 2386 | } |
2381 | 2387 | ||
2382 | void QRegExpEngine::Box::plus( int atom ) | 2388 | void QRegExpEngine::Box::plus( int atom ) |
2383 | { | 2389 | { |
2384 | #ifndef QT_NO_REGEXP_CAPTURE | 2390 | #ifndef QT_NO_REGEXP_CAPTURE |
2385 | eng->addPlusTransitions( rs, ls, atom ); | 2391 | eng->addPlusTransitions( rs, ls, atom ); |
2386 | #else | 2392 | #else |
2387 | Q_UNUSED( atom ); | 2393 | Q_UNUSED( atom ); |
2388 | eng->addCatTransitions( rs, ls ); | 2394 | eng->addCatTransitions( rs, ls ); |
2389 | #endif | 2395 | #endif |
2390 | addAnchorsToEngine( *this ); | 2396 | addAnchorsToEngine( *this ); |
2391 | #ifndef QT_NO_REGEXP_OPTIM | 2397 | #ifndef QT_NO_REGEXP_OPTIM |
2392 | maxl = InftyLen; | 2398 | maxl = InftyLen; |
2393 | #endif | 2399 | #endif |
2394 | } | 2400 | } |
2395 | 2401 | ||
2396 | void QRegExpEngine::Box::opt() | 2402 | void QRegExpEngine::Box::opt() |
2397 | { | 2403 | { |
2398 | #ifndef QT_NO_REGEXP_OPTIM | 2404 | #ifndef QT_NO_REGEXP_OPTIM |
2399 | earlyStart = 0; | 2405 | earlyStart = 0; |
2400 | lateStart = 0; | 2406 | lateStart = 0; |
2401 | str = QString::null; | 2407 | str = QString::null; |
2402 | leftStr = QString::null; | 2408 | leftStr = QString::null; |
2403 | rightStr = QString::null; | 2409 | rightStr = QString::null; |
2404 | #endif | 2410 | #endif |
2405 | skipanchors = 0; | 2411 | skipanchors = 0; |
2406 | minl = 0; | 2412 | minl = 0; |
2407 | } | 2413 | } |
2408 | 2414 | ||
2409 | void QRegExpEngine::Box::catAnchor( int a ) | 2415 | void QRegExpEngine::Box::catAnchor( int a ) |
2410 | { | 2416 | { |
2411 | if ( a != 0 ) { | 2417 | if ( a != 0 ) { |
2412 | for ( int i = 0; i < (int) rs.size(); i++ ) { | 2418 | for ( int i = 0; i < (int) rs.size(); i++ ) { |
2413 | a = eng->anchorConcatenation( at(ranchors, rs[i]), a ); | 2419 | a = eng->anchorConcatenation( at(ranchors, rs[i]), a ); |
2414 | ranchors.insert( rs[i], a ); | 2420 | ranchors.insert( rs[i], a ); |
2415 | } | 2421 | } |
2416 | if ( minl == 0 ) | 2422 | if ( minl == 0 ) |
2417 | skipanchors = eng->anchorConcatenation( skipanchors, a ); | 2423 | skipanchors = eng->anchorConcatenation( skipanchors, a ); |
2418 | } | 2424 | } |
2419 | } | 2425 | } |
2420 | 2426 | ||
2421 | #ifndef QT_NO_REGEXP_OPTIM | 2427 | #ifndef QT_NO_REGEXP_OPTIM |
2422 | void QRegExpEngine::Box::setupHeuristics() | 2428 | void QRegExpEngine::Box::setupHeuristics() |
2423 | { | 2429 | { |
2424 | eng->setupGoodStringHeuristic( earlyStart, lateStart, str ); | 2430 | eng->setupGoodStringHeuristic( earlyStart, lateStart, str ); |
2425 | 2431 | ||
2426 | /* | 2432 | /* |
2427 | A regular expression such as 112|1 has occ1['2'] = 2 and minl = 1 at this | 2433 | A regular expression such as 112|1 has occ1['2'] = 2 and minl = 1 at this |
2428 | point. An entry of occ1 has to be at most minl or infinity for the rest | 2434 | point. An entry of occ1 has to be at most minl or infinity for the rest |
2429 | of the algorithm to go well. | 2435 | of the algorithm to go well. |
2430 | 2436 | ||
2431 | We waited until here before normalizing these cases (instead of doing it | 2437 | We waited until here before normalizing these cases (instead of doing it |
2432 | in Box::orx()) because sometimes things improve by themselves; consider | 2438 | in Box::orx()) because sometimes things improve by themselves; consider |
2433 | (112|1)34. | 2439 | (112|1)34. |
2434 | */ | 2440 | */ |
2435 | for ( int i = 0; i < NumBadChars; i++ ) { | 2441 | for ( int i = 0; i < NumBadChars; i++ ) { |
2436 | if ( occ1[i] != NoOccurrence && occ1[i] >= minl ) | 2442 | if ( occ1[i] != NoOccurrence && occ1[i] >= minl ) |
2437 | occ1[i] = minl; | 2443 | occ1[i] = minl; |
2438 | } | 2444 | } |
2439 | eng->setupBadCharHeuristic( minl, occ1 ); | 2445 | eng->setupBadCharHeuristic( minl, occ1 ); |
2440 | 2446 | ||
2441 | eng->heuristicallyChooseHeuristic(); | 2447 | eng->heuristicallyChooseHeuristic(); |
2442 | } | 2448 | } |
2443 | #endif | 2449 | #endif |
2444 | 2450 | ||
2445 | #if defined(QT_DEBUG) | 2451 | #if defined(QT_DEBUG) |
2446 | void QRegExpEngine::Box::dump() const | 2452 | void QRegExpEngine::Box::dump() const |
2447 | { | 2453 | { |
2448 | int i; | 2454 | int i; |
2449 | qDebug( "Box of at least %d character%s", minl, minl == 1 ? "" : "s" ); | 2455 | odebug << "Box of at least " << minl << " character" << (minl == 1 ? "" : "s") << oendl; |
2450 | qDebug( " Left states:" ); | 2456 | odebug << " Left states:" << oendl; |
2451 | for ( i = 0; i < (int) ls.size(); i++ ) { | 2457 | for ( i = 0; i < (int) ls.size(); i++ ) { |
2452 | if ( at(lanchors, ls[i]) == 0 ) | 2458 | if ( at(lanchors, ls[i]) == 0 ) |
2453 | qDebug( " %d", ls[i] ); | 2459 | odebug << " " << ls[i] << oendl; |
2454 | else | 2460 | else |
2455 | qDebug( " %d [anchors 0x%.8x]", ls[i], lanchors[ls[i]] ); | 2461 | odebug << " " << ls[i] << QString().sprintf(" [anchors 0x%.8x]", lanchors[ls[i]]) << oendl; |
2456 | } | 2462 | } |
2457 | qDebug( " Right states:" ); | 2463 | odebug << " Right states:" << oendl; |
2458 | for ( i = 0; i < (int) rs.size(); i++ ) { | 2464 | for ( i = 0; i < (int) rs.size(); i++ ) { |
2459 | if ( at(ranchors, ls[i]) == 0 ) | 2465 | if ( at(ranchors, ls[i]) == 0 ) |
2460 | qDebug( " %d", rs[i] ); | 2466 | odebug << " " << rs[i] << oendl; |
2461 | else | 2467 | else |
2462 | qDebug( " %d [anchors 0x%.8x]", rs[i], ranchors[rs[i]] ); | 2468 | odebug << " " << rs[i] << QString().sprintf(" [anchors 0x%.8x]", ranchors[rs[i]]) << oendl; |
2463 | } | 2469 | } |
2464 | qDebug( " Skip anchors: 0x%.8x", skipanchors ); | 2470 | odebug << QString().sprintf(" Skip anchors: 0x%.8x", skipanchors) << oendl; |
2465 | } | 2471 | } |
2466 | #endif | 2472 | #endif |
2467 | 2473 | ||
2468 | void QRegExpEngine::Box::addAnchorsToEngine( const Box& to ) const | 2474 | void QRegExpEngine::Box::addAnchorsToEngine( const Box& to ) const |
2469 | { | 2475 | { |
2470 | for ( int i = 0; i < (int) to.ls.size(); i++ ) { | 2476 | for ( int i = 0; i < (int) to.ls.size(); i++ ) { |
2471 | for ( int j = 0; j < (int) rs.size(); j++ ) { | 2477 | for ( int j = 0; j < (int) rs.size(); j++ ) { |
2472 | int a = eng->anchorConcatenation( at(ranchors, rs[j]), | 2478 | int a = eng->anchorConcatenation( at(ranchors, rs[j]), |
2473 | at(to.lanchors, to.ls[i]) ); | 2479 | at(to.lanchors, to.ls[i]) ); |
2474 | eng->addAnchors( rs[j], to.ls[i], a ); | 2480 | eng->addAnchors( rs[j], to.ls[i], a ); |
2475 | } | 2481 | } |
2476 | } | 2482 | } |
2477 | } | 2483 | } |
2478 | 2484 | ||
2479 | int QRegExpEngine::getChar() | 2485 | int QRegExpEngine::getChar() |
2480 | { | 2486 | { |
2481 | return ( yyPos == yyLen ) ? EOS : yyIn[yyPos++].unicode(); | 2487 | return ( yyPos == yyLen ) ? EOS : yyIn[yyPos++].unicode(); |
2482 | } | 2488 | } |
2483 | 2489 | ||
2484 | int QRegExpEngine::getEscape() | 2490 | int QRegExpEngine::getEscape() |
2485 | { | 2491 | { |
2486 | #ifndef QT_NO_REGEXP_ESCAPE | 2492 | #ifndef QT_NO_REGEXP_ESCAPE |
2487 | const char tab[] = "afnrtv"; // no b, as \b means word boundary | 2493 | const char tab[] = "afnrtv"; // no b, as \b means word boundary |
2488 | const char backTab[] = "\a\f\n\r\t\v"; | 2494 | const char backTab[] = "\a\f\n\r\t\v"; |
2489 | ushort low; | 2495 | ushort low; |
2490 | int i; | 2496 | int i; |
2491 | #endif | 2497 | #endif |
2492 | ushort val; | 2498 | ushort val; |
2493 | int prevCh = yyCh; | 2499 | int prevCh = yyCh; |
2494 | 2500 | ||
2495 | if ( prevCh == EOS ) { | 2501 | if ( prevCh == EOS ) { |
2496 | yyError = TRUE; | 2502 | yyError = TRUE; |
2497 | return Tok_Char | '\\'; | 2503 | return Tok_Char | '\\'; |
2498 | } | 2504 | } |
2499 | yyCh = getChar(); | 2505 | yyCh = getChar(); |
2500 | #ifndef QT_NO_REGEXP_ESCAPE | 2506 | #ifndef QT_NO_REGEXP_ESCAPE |
2501 | if ( (prevCh & ~0xff) == 0 ) { | 2507 | if ( (prevCh & ~0xff) == 0 ) { |
2502 | const char *p = strchr( tab, prevCh ); | 2508 | const char *p = strchr( tab, prevCh ); |
2503 | if ( p != 0 ) | 2509 | if ( p != 0 ) |
2504 | return Tok_Char | backTab[p - tab]; | 2510 | return Tok_Char | backTab[p - tab]; |
2505 | } | 2511 | } |
2506 | #endif | 2512 | #endif |
2507 | 2513 | ||
2508 | switch ( prevCh ) { | 2514 | switch ( prevCh ) { |
2509 | #ifndef QT_NO_REGEXP_ESCAPE | 2515 | #ifndef QT_NO_REGEXP_ESCAPE |
2510 | case '0': | 2516 | case '0': |
2511 | val = 0; | 2517 | val = 0; |
2512 | for ( i = 0; i < 3; i++ ) { | 2518 | for ( i = 0; i < 3; i++ ) { |
2513 | if ( yyCh >= '0' && yyCh <= '7' ) | 2519 | if ( yyCh >= '0' && yyCh <= '7' ) |
2514 | val = ( val << 3 ) | ( yyCh - '0' ); | 2520 | val = ( val << 3 ) | ( yyCh - '0' ); |
2515 | else | 2521 | else |
2516 | break; | 2522 | break; |
2517 | yyCh = getChar(); | 2523 | yyCh = getChar(); |
2518 | } | 2524 | } |
2519 | if ( (val & ~0377) != 0 ) | 2525 | if ( (val & ~0377) != 0 ) |
2520 | yyError = TRUE; | 2526 | yyError = TRUE; |
2521 | return Tok_Char | val; | 2527 | return Tok_Char | val; |
2522 | #endif | 2528 | #endif |
2523 | #ifndef QT_NO_REGEXP_ESCAPE | 2529 | #ifndef QT_NO_REGEXP_ESCAPE |
2524 | case 'B': | 2530 | case 'B': |
2525 | return Tok_NonWord; | 2531 | return Tok_NonWord; |
2526 | #endif | 2532 | #endif |
2527 | #ifndef QT_NO_REGEXP_CCLASS | 2533 | #ifndef QT_NO_REGEXP_CCLASS |
2528 | case 'D': | 2534 | case 'D': |
2529 | // see QChar::isDigit() | 2535 | // see QChar::isDigit() |
2530 | yyCharClass->addCategories( 0x7fffffef ); | 2536 | yyCharClass->addCategories( 0x7fffffef ); |
2531 | return Tok_CharClass; | 2537 | return Tok_CharClass; |
2532 | case 'S': | 2538 | case 'S': |
2533 | // see QChar::isSpace() | 2539 | // see QChar::isSpace() |
2534 | yyCharClass->addCategories( 0x7ffff87f ); | 2540 | yyCharClass->addCategories( 0x7ffff87f ); |
2535 | yyCharClass->addRange( 0x0000, 0x0008 ); | 2541 | yyCharClass->addRange( 0x0000, 0x0008 ); |
2536 | yyCharClass->addRange( 0x000e, 0x001f ); | 2542 | yyCharClass->addRange( 0x000e, 0x001f ); |
2537 | yyCharClass->addRange( 0x007f, 0x009f ); | 2543 | yyCharClass->addRange( 0x007f, 0x009f ); |
2538 | return Tok_CharClass; | 2544 | return Tok_CharClass; |
2539 | case 'W': | 2545 | case 'W': |
2540 | // see QChar::isLetterOrNumber() | 2546 | // see QChar::isLetterOrNumber() |
2541 | yyCharClass->addCategories( 0x7ff07f8f ); | 2547 | yyCharClass->addCategories( 0x7ff07f8f ); |
2542 | return Tok_CharClass; | 2548 | return Tok_CharClass; |
2543 | #endif | 2549 | #endif |
2544 | #ifndef QT_NO_REGEXP_ESCAPE | 2550 | #ifndef QT_NO_REGEXP_ESCAPE |
2545 | case 'b': | 2551 | case 'b': |
2546 | return Tok_Word; | 2552 | return Tok_Word; |
2547 | #endif | 2553 | #endif |
2548 | #ifndef QT_NO_REGEXP_CCLASS | 2554 | #ifndef QT_NO_REGEXP_CCLASS |
2549 | case 'd': | 2555 | case 'd': |
2550 | // see QChar::isDigit() | 2556 | // see QChar::isDigit() |
2551 | yyCharClass->addCategories( 0x00000010 ); | 2557 | yyCharClass->addCategories( 0x00000010 ); |
2552 | return Tok_CharClass; | 2558 | return Tok_CharClass; |
2553 | case 's': | 2559 | case 's': |
2554 | // see QChar::isSpace() | 2560 | // see QChar::isSpace() |
2555 | yyCharClass->addCategories( 0x00000380 ); | 2561 | yyCharClass->addCategories( 0x00000380 ); |
2556 | yyCharClass->addRange( 0x0009, 0x000d ); | 2562 | yyCharClass->addRange( 0x0009, 0x000d ); |
2557 | return Tok_CharClass; | 2563 | return Tok_CharClass; |
2558 | case 'w': | 2564 | case 'w': |
2559 | // see QChar::isLetterOrNumber() | 2565 | // see QChar::isLetterOrNumber() |
2560 | yyCharClass->addCategories( 0x000f8070 ); | 2566 | yyCharClass->addCategories( 0x000f8070 ); |
2561 | return Tok_CharClass; | 2567 | return Tok_CharClass; |
2562 | #endif | 2568 | #endif |
2563 | #ifndef QT_NO_REGEXP_ESCAPE | 2569 | #ifndef QT_NO_REGEXP_ESCAPE |
2564 | case 'x': | 2570 | case 'x': |
2565 | val = 0; | 2571 | val = 0; |
2566 | for ( i = 0; i < 4; i++ ) { | 2572 | for ( i = 0; i < 4; i++ ) { |
2567 | low = QChar( yyCh ).lower(); | 2573 | low = QChar( yyCh ).lower(); |
2568 | if ( low >= '0' && low <= '9' ) | 2574 | if ( low >= '0' && low <= '9' ) |
2569 | val = ( val << 4 ) | ( low - '0' ); | 2575 | val = ( val << 4 ) | ( low - '0' ); |
2570 | else if ( low >= 'a' && low <= 'f' ) | 2576 | else if ( low >= 'a' && low <= 'f' ) |
2571 | val = ( val << 4 ) | ( low - 'a' + 10 ); | 2577 | val = ( val << 4 ) | ( low - 'a' + 10 ); |
2572 | else | 2578 | else |
2573 | break; | 2579 | break; |
2574 | yyCh = getChar(); | 2580 | yyCh = getChar(); |
2575 | } | 2581 | } |
2576 | return Tok_Char | val; | 2582 | return Tok_Char | val; |
2577 | #endif | 2583 | #endif |
2578 | default: | 2584 | default: |
2579 | if ( prevCh >= '1' && prevCh <= '9' ) { | 2585 | if ( prevCh >= '1' && prevCh <= '9' ) { |
2580 | #ifndef QT_NO_REGEXP_BACKREF | 2586 | #ifndef QT_NO_REGEXP_BACKREF |
2581 | val = prevCh - '0'; | 2587 | val = prevCh - '0'; |
2582 | while ( yyCh >= '0' && yyCh <= '9' ) { | 2588 | while ( yyCh >= '0' && yyCh <= '9' ) { |
2583 | val = ( val *= 10 ) | ( yyCh - '0' ); | 2589 | val = ( val *= 10 ) | ( yyCh - '0' ); |
2584 | yyCh = getChar(); | 2590 | yyCh = getChar(); |
2585 | } | 2591 | } |
2586 | return Tok_BackRef | val; | 2592 | return Tok_BackRef | val; |
2587 | #else | 2593 | #else |
2588 | yyError = TRUE; | 2594 | yyError = TRUE; |
2589 | #endif | 2595 | #endif |
2590 | } | 2596 | } |
2591 | return Tok_Char | prevCh; | 2597 | return Tok_Char | prevCh; |
2592 | } | 2598 | } |
2593 | } | 2599 | } |
2594 | 2600 | ||
2595 | #ifndef QT_NO_REGEXP_INTERVAL | 2601 | #ifndef QT_NO_REGEXP_INTERVAL |
2596 | int QRegExpEngine::getRep( int def ) | 2602 | int QRegExpEngine::getRep( int def ) |
2597 | { | 2603 | { |
2598 | if ( yyCh >= '0' && yyCh <= '9' ) { | 2604 | if ( yyCh >= '0' && yyCh <= '9' ) { |
2599 | int rep = 0; | 2605 | int rep = 0; |
2600 | do { | 2606 | do { |
2601 | rep = 10 * rep + yyCh - '0'; | 2607 | rep = 10 * rep + yyCh - '0'; |
2602 | if ( rep >= InftyRep ) { | 2608 | if ( rep >= InftyRep ) { |
2603 | yyError = TRUE; | 2609 | yyError = TRUE; |
2604 | rep = def; | 2610 | rep = def; |
2605 | } | 2611 | } |
2606 | yyCh = getChar(); | 2612 | yyCh = getChar(); |
2607 | } while ( yyCh >= '0' && yyCh <= '9' ); | 2613 | } while ( yyCh >= '0' && yyCh <= '9' ); |
2608 | return rep; | 2614 | return rep; |
2609 | } else { | 2615 | } else { |
2610 | return def; | 2616 | return def; |
2611 | } | 2617 | } |
2612 | } | 2618 | } |
2613 | #endif | 2619 | #endif |
2614 | 2620 | ||
2615 | #ifndef QT_NO_REGEXP_LOOKAHEAD | 2621 | #ifndef QT_NO_REGEXP_LOOKAHEAD |
2616 | void QRegExpEngine::skipChars( int n ) | 2622 | void QRegExpEngine::skipChars( int n ) |
2617 | { | 2623 | { |
2618 | if ( n > 0 ) { | 2624 | if ( n > 0 ) { |
2619 | yyPos += n - 1; | 2625 | yyPos += n - 1; |
2620 | yyCh = getChar(); | 2626 | yyCh = getChar(); |
2621 | } | 2627 | } |
2622 | } | 2628 | } |
2623 | #endif | 2629 | #endif |
2624 | 2630 | ||
2625 | void QRegExpEngine::startTokenizer( const QChar *rx, int len ) | 2631 | void QRegExpEngine::startTokenizer( const QChar *rx, int len ) |
2626 | { | 2632 | { |
2627 | yyIn = rx; | 2633 | yyIn = rx; |
2628 | yyPos0 = 0; | 2634 | yyPos0 = 0; |
2629 | yyPos = 0; | 2635 | yyPos = 0; |
2630 | yyLen = len; | 2636 | yyLen = len; |
2631 | yyCh = getChar(); | 2637 | yyCh = getChar(); |
2632 | yyCharClass = new CharClass; | 2638 | yyCharClass = new CharClass; |
2633 | yyMinRep = 0; | 2639 | yyMinRep = 0; |
2634 | yyMaxRep = 0; | 2640 | yyMaxRep = 0; |
2635 | yyError = FALSE; | 2641 | yyError = FALSE; |
2636 | } | 2642 | } |
2637 | 2643 | ||
2638 | int QRegExpEngine::getToken() | 2644 | int QRegExpEngine::getToken() |
2639 | { | 2645 | { |
2640 | #ifndef QT_NO_REGEXP_CCLASS | 2646 | #ifndef QT_NO_REGEXP_CCLASS |
2641 | ushort pendingCh = 0; | 2647 | ushort pendingCh = 0; |
2642 | bool charPending; | 2648 | bool charPending; |
2643 | bool rangePending; | 2649 | bool rangePending; |
2644 | int tok; | 2650 | int tok; |
2645 | #endif | 2651 | #endif |
2646 | int prevCh = yyCh; | 2652 | int prevCh = yyCh; |
2647 | 2653 | ||
2648 | yyPos0 = yyPos - 1; | 2654 | yyPos0 = yyPos - 1; |
2649 | #ifndef QT_NO_REGEXP_CCLASS | 2655 | #ifndef QT_NO_REGEXP_CCLASS |
2650 | yyCharClass->clear(); | 2656 | yyCharClass->clear(); |
2651 | #endif | 2657 | #endif |
2652 | yyMinRep = 0; | 2658 | yyMinRep = 0; |
2653 | yyMaxRep = 0; | 2659 | yyMaxRep = 0; |
2654 | yyCh = getChar(); | 2660 | yyCh = getChar(); |
2655 | switch ( prevCh ) { | 2661 | switch ( prevCh ) { |
2656 | case EOS: | 2662 | case EOS: |
2657 | yyPos0 = yyPos; | 2663 | yyPos0 = yyPos; |
2658 | return Tok_Eos; | 2664 | return Tok_Eos; |
2659 | case '$': | 2665 | case '$': |
2660 | return Tok_Dollar; | 2666 | return Tok_Dollar; |
2661 | case '(': | 2667 | case '(': |
2662 | if ( yyCh == '?' ) { | 2668 | if ( yyCh == '?' ) { |
2663 | prevCh = getChar(); | 2669 | prevCh = getChar(); |
2664 | yyCh = getChar(); | 2670 | yyCh = getChar(); |
2665 | switch ( prevCh ) { | 2671 | switch ( prevCh ) { |
2666 | #ifndef QT_NO_REGEXP_LOOKAHEAD | 2672 | #ifndef QT_NO_REGEXP_LOOKAHEAD |
2667 | case '!': | 2673 | case '!': |
2668 | return Tok_NegLookahead; | 2674 | return Tok_NegLookahead; |
2669 | case '=': | 2675 | case '=': |
2670 | return Tok_PosLookahead; | 2676 | return Tok_PosLookahead; |
2671 | #endif | 2677 | #endif |
2672 | case ':': | 2678 | case ':': |
2673 | return Tok_MagicLeftParen; | 2679 | return Tok_MagicLeftParen; |
2674 | default: | 2680 | default: |
2675 | yyError = TRUE; | 2681 | yyError = TRUE; |
2676 | return Tok_MagicLeftParen; | 2682 | return Tok_MagicLeftParen; |
2677 | } | 2683 | } |
2678 | } else { | 2684 | } else { |
2679 | return Tok_LeftParen; | 2685 | return Tok_LeftParen; |
2680 | } | 2686 | } |
2681 | case ')': | 2687 | case ')': |
2682 | return Tok_RightParen; | 2688 | return Tok_RightParen; |
2683 | case '*': | 2689 | case '*': |
2684 | yyMinRep = 0; | 2690 | yyMinRep = 0; |
2685 | yyMaxRep = InftyRep; | 2691 | yyMaxRep = InftyRep; |
2686 | return Tok_Quantifier; | 2692 | return Tok_Quantifier; |
2687 | case '+': | 2693 | case '+': |
2688 | yyMinRep = 1; | 2694 | yyMinRep = 1; |
2689 | yyMaxRep = InftyRep; | 2695 | yyMaxRep = InftyRep; |
2690 | return Tok_Quantifier; | 2696 | return Tok_Quantifier; |
2691 | case '.': | 2697 | case '.': |
2692 | #ifndef QT_NO_REGEXP_CCLASS | 2698 | #ifndef QT_NO_REGEXP_CCLASS |
2693 | yyCharClass->setNegative( TRUE ); | 2699 | yyCharClass->setNegative( TRUE ); |
2694 | #endif | 2700 | #endif |
2695 | return Tok_CharClass; | 2701 | return Tok_CharClass; |
2696 | case '?': | 2702 | case '?': |
2697 | yyMinRep = 0; | 2703 | yyMinRep = 0; |
2698 | yyMaxRep = 1; | 2704 | yyMaxRep = 1; |
2699 | return Tok_Quantifier; | 2705 | return Tok_Quantifier; |
2700 | case '[': | 2706 | case '[': |
2701 | #ifndef QT_NO_REGEXP_CCLASS | 2707 | #ifndef QT_NO_REGEXP_CCLASS |
2702 | if ( yyCh == '^' ) { | 2708 | if ( yyCh == '^' ) { |
2703 | yyCharClass->setNegative( TRUE ); | 2709 | yyCharClass->setNegative( TRUE ); |
2704 | yyCh = getChar(); | 2710 | yyCh = getChar(); |
2705 | } | 2711 | } |
2706 | charPending = FALSE; | 2712 | charPending = FALSE; |
2707 | rangePending = FALSE; | 2713 | rangePending = FALSE; |
2708 | do { | 2714 | do { |
2709 | if ( yyCh == '-' && charPending && !rangePending ) { | 2715 | if ( yyCh == '-' && charPending && !rangePending ) { |
2710 | rangePending = TRUE; | 2716 | rangePending = TRUE; |
2711 | yyCh = getChar(); | 2717 | yyCh = getChar(); |
2712 | } else { | 2718 | } else { |
2713 | if ( charPending && !rangePending ) { | 2719 | if ( charPending && !rangePending ) { |
2714 | yyCharClass->addSingleton( pendingCh ); | 2720 | yyCharClass->addSingleton( pendingCh ); |
2715 | charPending = FALSE; | 2721 | charPending = FALSE; |
2716 | } | 2722 | } |
2717 | if ( yyCh == '\\' ) { | 2723 | if ( yyCh == '\\' ) { |
2718 | yyCh = getChar(); | 2724 | yyCh = getChar(); |
2719 | tok = getEscape(); | 2725 | tok = getEscape(); |
2720 | if ( tok == Tok_Word ) | 2726 | if ( tok == Tok_Word ) |
2721 | tok = '\b'; | 2727 | tok = '\b'; |
2722 | } else { | 2728 | } else { |
2723 | tok = Tok_Char | yyCh; | 2729 | tok = Tok_Char | yyCh; |
2724 | yyCh = getChar(); | 2730 | yyCh = getChar(); |
2725 | } | 2731 | } |
2726 | if ( tok == Tok_CharClass ) { | 2732 | if ( tok == Tok_CharClass ) { |
2727 | if ( rangePending ) { | 2733 | if ( rangePending ) { |
2728 | yyCharClass->addSingleton( '-' ); | 2734 | yyCharClass->addSingleton( '-' ); |
2729 | yyCharClass->addSingleton( pendingCh ); | 2735 | yyCharClass->addSingleton( pendingCh ); |
2730 | charPending = FALSE; | 2736 | charPending = FALSE; |
2731 | rangePending = FALSE; | 2737 | rangePending = FALSE; |
2732 | } | 2738 | } |
2733 | } else if ( (tok & Tok_Char) != 0 ) { | 2739 | } else if ( (tok & Tok_Char) != 0 ) { |
2734 | if ( rangePending ) { | 2740 | if ( rangePending ) { |
2735 | yyCharClass->addRange( pendingCh, tok ^ Tok_Char ); | 2741 | yyCharClass->addRange( pendingCh, tok ^ Tok_Char ); |
2736 | charPending = FALSE; | 2742 | charPending = FALSE; |
2737 | rangePending = FALSE; | 2743 | rangePending = FALSE; |
2738 | } else { | 2744 | } else { |
2739 | pendingCh = tok ^ Tok_Char; | 2745 | pendingCh = tok ^ Tok_Char; |
2740 | charPending = TRUE; | 2746 | charPending = TRUE; |
2741 | } | 2747 | } |
2742 | } else { | 2748 | } else { |
2743 | yyError = TRUE; | 2749 | yyError = TRUE; |
2744 | } | 2750 | } |
2745 | } | 2751 | } |
2746 | } while ( yyCh != ']' && yyCh != EOS ); | 2752 | } while ( yyCh != ']' && yyCh != EOS ); |
2747 | if ( rangePending ) | 2753 | if ( rangePending ) |
2748 | yyCharClass->addSingleton( '-' ); | 2754 | yyCharClass->addSingleton( '-' ); |
2749 | if ( charPending ) | 2755 | if ( charPending ) |
2750 | yyCharClass->addSingleton( pendingCh ); | 2756 | yyCharClass->addSingleton( pendingCh ); |
2751 | if ( yyCh == EOS ) | 2757 | if ( yyCh == EOS ) |
2752 | yyError = TRUE; | 2758 | yyError = TRUE; |
2753 | else | 2759 | else |
2754 | yyCh = getChar(); | 2760 | yyCh = getChar(); |
2755 | return Tok_CharClass; | 2761 | return Tok_CharClass; |
2756 | #else | 2762 | #else |
2757 | yyError = TRUE; | 2763 | yyError = TRUE; |
2758 | return Tok_Char | '['; | 2764 | return Tok_Char | '['; |
2759 | #endif | 2765 | #endif |
2760 | case '\\': | 2766 | case '\\': |
2761 | return getEscape(); | 2767 | return getEscape(); |
2762 | case ']': | 2768 | case ']': |
2763 | yyError = TRUE; | 2769 | yyError = TRUE; |
2764 | return Tok_Char | ']'; | 2770 | return Tok_Char | ']'; |
2765 | case '^': | 2771 | case '^': |
2766 | return Tok_Caret; | 2772 | return Tok_Caret; |
2767 | #ifndef QT_NO_REGEXP_INTERVAL | 2773 | #ifndef QT_NO_REGEXP_INTERVAL |
2768 | case '{': | 2774 | case '{': |
2769 | yyMinRep = getRep( 0 ); | 2775 | yyMinRep = getRep( 0 ); |
2770 | yyMaxRep = yyMinRep; | 2776 | yyMaxRep = yyMinRep; |
2771 | if ( yyCh == ',' ) { | 2777 | if ( yyCh == ',' ) { |
2772 | yyCh = getChar(); | 2778 | yyCh = getChar(); |
2773 | yyMaxRep = getRep( InftyRep ); | 2779 | yyMaxRep = getRep( InftyRep ); |
2774 | } | 2780 | } |
2775 | if ( yyMaxRep < yyMinRep ) | 2781 | if ( yyMaxRep < yyMinRep ) |
2776 | qSwap( yyMinRep, yyMaxRep ); | 2782 | qSwap( yyMinRep, yyMaxRep ); |
2777 | if ( yyCh != '}' ) | 2783 | if ( yyCh != '}' ) |
2778 | yyError = TRUE; | 2784 | yyError = TRUE; |
2779 | yyCh = getChar(); | 2785 | yyCh = getChar(); |
2780 | return Tok_Quantifier; | 2786 | return Tok_Quantifier; |
2781 | #else | 2787 | #else |
2782 | yyError = TRUE; | 2788 | yyError = TRUE; |
2783 | return Tok_Char | '{'; | 2789 | return Tok_Char | '{'; |
2784 | #endif | 2790 | #endif |
2785 | case '|': | 2791 | case '|': |
2786 | return Tok_Bar; | 2792 | return Tok_Bar; |
2787 | case '}': | 2793 | case '}': |
2788 | yyError = TRUE; | 2794 | yyError = TRUE; |
2789 | return Tok_Char | '}'; | 2795 | return Tok_Char | '}'; |
2790 | default: | 2796 | default: |
2791 | return Tok_Char | prevCh; | 2797 | return Tok_Char | prevCh; |
2792 | } | 2798 | } |
2793 | } | 2799 | } |
2794 | 2800 | ||
2795 | int QRegExpEngine::parse( const QChar *pattern, int len ) | 2801 | int QRegExpEngine::parse( const QChar *pattern, int len ) |
2796 | { | 2802 | { |
2797 | valid = TRUE; | 2803 | valid = TRUE; |
2798 | startTokenizer( pattern, len ); | 2804 | startTokenizer( pattern, len ); |
2799 | yyTok = getToken(); | 2805 | yyTok = getToken(); |
2800 | #ifndef QT_NO_REGEXP_CAPTURE | 2806 | #ifndef QT_NO_REGEXP_CAPTURE |
2801 | yyMayCapture = TRUE; | 2807 | yyMayCapture = TRUE; |
2802 | #else | 2808 | #else |
2803 | yyMayCapture = FALSE; | 2809 | yyMayCapture = FALSE; |
2804 | #endif | 2810 | #endif |
2805 | 2811 | ||
2806 | #ifndef QT_NO_REGEXP_CAPTURE | 2812 | #ifndef QT_NO_REGEXP_CAPTURE |
2807 | int atom = startAtom( FALSE ); | 2813 | int atom = startAtom( FALSE ); |
2808 | #endif | 2814 | #endif |
2809 | CharClass anything; | 2815 | CharClass anything; |
2810 | Box box( this ); // create InitialState | 2816 | Box box( this ); // create InitialState |
2811 | box.set( anything ); | 2817 | box.set( anything ); |
2812 | Box rightBox( this ); // create FinalState | 2818 | Box rightBox( this ); // create FinalState |
2813 | rightBox.set( anything ); | 2819 | rightBox.set( anything ); |
2814 | 2820 | ||
2815 | Box middleBox( this ); | 2821 | Box middleBox( this ); |
2816 | parseExpression( &middleBox ); | 2822 | parseExpression( &middleBox ); |
2817 | #ifndef QT_NO_REGEXP_CAPTURE | 2823 | #ifndef QT_NO_REGEXP_CAPTURE |
2818 | finishAtom( atom ); | 2824 | finishAtom( atom ); |
2819 | #endif | 2825 | #endif |
2820 | #ifndef QT_NO_REGEXP_OPTIM | 2826 | #ifndef QT_NO_REGEXP_OPTIM |
2821 | middleBox.setupHeuristics(); | 2827 | middleBox.setupHeuristics(); |
2822 | #endif | 2828 | #endif |
2823 | box.cat( middleBox ); | 2829 | box.cat( middleBox ); |
2824 | box.cat( rightBox ); | 2830 | box.cat( rightBox ); |
2825 | delete yyCharClass; | 2831 | delete yyCharClass; |
2826 | yyCharClass = 0; | 2832 | yyCharClass = 0; |
2827 | 2833 | ||
2828 | realncap = ncap; | 2834 | realncap = ncap; |
2829 | #ifndef QT_NO_REGEXP_BACKREF | 2835 | #ifndef QT_NO_REGEXP_BACKREF |
2830 | if ( nbrefs > ncap ) | 2836 | if ( nbrefs > ncap ) |
2831 | ncap = nbrefs; | 2837 | ncap = nbrefs; |
2832 | #endif | 2838 | #endif |
2833 | 2839 | ||
2834 | mmCaptured.resize( 2 + 2 * realncap ); | 2840 | mmCaptured.resize( 2 + 2 * realncap ); |
2835 | mmCapturedNoMatch.fill( -1, 2 + 2 * realncap ); | 2841 | mmCapturedNoMatch.fill( -1, 2 + 2 * realncap ); |
2836 | 2842 | ||
2837 | /* | 2843 | /* |
2838 | We use one QArray<int> for all the big data used a lot in matchHere() and | 2844 | We use one QArray<int> for all the big data used a lot in matchHere() and |
2839 | friends. | 2845 | friends. |
2840 | */ | 2846 | */ |
2841 | #ifndef QT_NO_REGEXP_OPTIM | 2847 | #ifndef QT_NO_REGEXP_OPTIM |
2842 | mmSlideTabSize = QMAX( minl + 1, 16 ); | 2848 | mmSlideTabSize = QMAX( minl + 1, 16 ); |
2843 | #else | 2849 | #else |
2844 | mmSlideTabSize = 0; | 2850 | mmSlideTabSize = 0; |
2845 | #endif | 2851 | #endif |
2846 | mmBigArray.resize( (3 + 4 * ncap) * ns + 4 * ncap + mmSlideTabSize ); | 2852 | mmBigArray.resize( (3 + 4 * ncap) * ns + 4 * ncap + mmSlideTabSize ); |
2847 | 2853 | ||
2848 | mmInNextStack = mmBigArray.data(); | 2854 | mmInNextStack = mmBigArray.data(); |
2849 | memset( mmInNextStack, -1, ns * sizeof(int) ); | 2855 | memset( mmInNextStack, -1, ns * sizeof(int) ); |
2850 | mmCurStack = mmInNextStack + ns; | 2856 | mmCurStack = mmInNextStack + ns; |
2851 | mmNextStack = mmInNextStack + 2 * ns; | 2857 | mmNextStack = mmInNextStack + 2 * ns; |
2852 | 2858 | ||
2853 | mmCurCapBegin = mmInNextStack + 3 * ns; | 2859 | mmCurCapBegin = mmInNextStack + 3 * ns; |
2854 | mmNextCapBegin = mmCurCapBegin + ncap * ns; | 2860 | mmNextCapBegin = mmCurCapBegin + ncap * ns; |
2855 | mmCurCapEnd = mmCurCapBegin + 2 * ncap * ns; | 2861 | mmCurCapEnd = mmCurCapBegin + 2 * ncap * ns; |
2856 | mmNextCapEnd = mmCurCapBegin + 3 * ncap * ns; | 2862 | mmNextCapEnd = mmCurCapBegin + 3 * ncap * ns; |
2857 | 2863 | ||
2858 | mmTempCapBegin = mmCurCapBegin + 4 * ncap * ns; | 2864 | mmTempCapBegin = mmCurCapBegin + 4 * ncap * ns; |
2859 | mmTempCapEnd = mmTempCapBegin + ncap; | 2865 | mmTempCapEnd = mmTempCapBegin + ncap; |
2860 | mmCapBegin = mmTempCapBegin + 2 * ncap; | 2866 | mmCapBegin = mmTempCapBegin + 2 * ncap; |
2861 | mmCapEnd = mmTempCapBegin + 3 * ncap; | 2867 | mmCapEnd = mmTempCapBegin + 3 * ncap; |
2862 | 2868 | ||
2863 | mmSlideTab = mmTempCapBegin + 4 * ncap; | 2869 | mmSlideTab = mmTempCapBegin + 4 * ncap; |
2864 | 2870 | ||
2865 | if ( yyError ) | 2871 | if ( yyError ) |
2866 | return -1; | 2872 | return -1; |
2867 | 2873 | ||
2868 | #ifndef QT_NO_REGEXP_OPTIM | 2874 | #ifndef QT_NO_REGEXP_OPTIM |
2869 | State *sinit = s[InitialState]; | 2875 | State *sinit = s[InitialState]; |
2870 | caretAnchored = ( sinit->anchors != 0 ); | 2876 | caretAnchored = ( sinit->anchors != 0 ); |
2871 | if ( caretAnchored ) { | 2877 | if ( caretAnchored ) { |
2872 | QMap<int, int>& anchors = *sinit->anchors; | 2878 | QMap<int, int>& anchors = *sinit->anchors; |
2873 | QMap<int, int>::ConstIterator a; | 2879 | QMap<int, int>::ConstIterator a; |
2874 | for ( a = anchors.begin(); a != anchors.end(); ++a ) { | 2880 | for ( a = anchors.begin(); a != anchors.end(); ++a ) { |
2875 | #ifndef QT_NO_REGEXP_ANCHOR_ALT | 2881 | #ifndef QT_NO_REGEXP_ANCHOR_ALT |
2876 | if ( (*a & Anchor_Alternation) != 0 ) | 2882 | if ( (*a & Anchor_Alternation) != 0 ) |
2877 | break; | 2883 | break; |
2878 | #endif | 2884 | #endif |
2879 | if ( (*a & Anchor_Caret) == 0 ) { | 2885 | if ( (*a & Anchor_Caret) == 0 ) { |
2880 | caretAnchored = FALSE; | 2886 | caretAnchored = FALSE; |
2881 | break; | 2887 | break; |
2882 | } | 2888 | } |
2883 | } | 2889 | } |
2884 | } | 2890 | } |
2885 | #endif | 2891 | #endif |
2886 | return yyPos0; | 2892 | return yyPos0; |
2887 | } | 2893 | } |
2888 | 2894 | ||
2889 | void QRegExpEngine::parseAtom( Box *box ) | 2895 | void QRegExpEngine::parseAtom( Box *box ) |
2890 | { | 2896 | { |
2891 | #ifndef QT_NO_REGEXP_LOOKAHEAD | 2897 | #ifndef QT_NO_REGEXP_LOOKAHEAD |
2892 | QRegExpEngine *eng = 0; | 2898 | QRegExpEngine *eng = 0; |
2893 | bool neg; | 2899 | bool neg; |
2894 | int len; | 2900 | int len; |
2895 | #endif | 2901 | #endif |
2896 | 2902 | ||
2897 | switch ( yyTok ) { | 2903 | switch ( yyTok ) { |
2898 | case Tok_Dollar: | 2904 | case Tok_Dollar: |
2899 | box->catAnchor( Anchor_Dollar ); | 2905 | box->catAnchor( Anchor_Dollar ); |
2900 | break; | 2906 | break; |
2901 | case Tok_Caret: | 2907 | case Tok_Caret: |
2902 | box->catAnchor( Anchor_Caret ); | 2908 | box->catAnchor( Anchor_Caret ); |
2903 | break; | 2909 | break; |
2904 | #ifndef QT_NO_REGEXP_LOOKAHEAD | 2910 | #ifndef QT_NO_REGEXP_LOOKAHEAD |
2905 | case Tok_PosLookahead: | 2911 | case Tok_PosLookahead: |
2906 | case Tok_NegLookahead: | 2912 | case Tok_NegLookahead: |
2907 | neg = ( yyTok == Tok_NegLookahead ); | 2913 | neg = ( yyTok == Tok_NegLookahead ); |
2908 | eng = new QRegExpEngine( cs ); | 2914 | eng = new QRegExpEngine( cs ); |
2909 | len = eng->parse( yyIn + yyPos - 1, yyLen - yyPos + 1 ); | 2915 | len = eng->parse( yyIn + yyPos - 1, yyLen - yyPos + 1 ); |
2910 | if ( len >= 0 ) | 2916 | if ( len >= 0 ) |
2911 | skipChars( len ); | 2917 | skipChars( len ); |
2912 | else | 2918 | else |
2913 | yyError = TRUE; | 2919 | yyError = TRUE; |
2914 | box->catAnchor( addLookahead(eng, neg) ); | 2920 | box->catAnchor( addLookahead(eng, neg) ); |
2915 | yyTok = getToken(); | 2921 | yyTok = getToken(); |
2916 | if ( yyTok != Tok_RightParen ) | 2922 | if ( yyTok != Tok_RightParen ) |
2917 | yyError = TRUE; | 2923 | yyError = TRUE; |
2918 | break; | 2924 | break; |
2919 | #endif | 2925 | #endif |
2920 | #ifndef QT_NO_REGEXP_ESCAPE | 2926 | #ifndef QT_NO_REGEXP_ESCAPE |
2921 | case Tok_Word: | 2927 | case Tok_Word: |
2922 | box->catAnchor( Anchor_Word ); | 2928 | box->catAnchor( Anchor_Word ); |
2923 | break; | 2929 | break; |
2924 | case Tok_NonWord: | 2930 | case Tok_NonWord: |
2925 | box->catAnchor( Anchor_NonWord ); | 2931 | box->catAnchor( Anchor_NonWord ); |
2926 | break; | 2932 | break; |
2927 | #endif | 2933 | #endif |
2928 | case Tok_LeftParen: | 2934 | case Tok_LeftParen: |
2929 | case Tok_MagicLeftParen: | 2935 | case Tok_MagicLeftParen: |
2930 | yyTok = getToken(); | 2936 | yyTok = getToken(); |
2931 | parseExpression( box ); | 2937 | parseExpression( box ); |
2932 | if ( yyTok != Tok_RightParen ) | 2938 | if ( yyTok != Tok_RightParen ) |
2933 | yyError = TRUE; | 2939 | yyError = TRUE; |
2934 | break; | 2940 | break; |
2935 | case Tok_CharClass: | 2941 | case Tok_CharClass: |
2936 | box->set( *yyCharClass ); | 2942 | box->set( *yyCharClass ); |
2937 | break; | 2943 | break; |
2938 | default: | 2944 | default: |
2939 | if ( (yyTok & Tok_Char) != 0 ) | 2945 | if ( (yyTok & Tok_Char) != 0 ) |
2940 | box->set( QChar(yyTok ^ Tok_Char) ); | 2946 | box->set( QChar(yyTok ^ Tok_Char) ); |
2941 | #ifndef QT_NO_REGEXP_BACKREF | 2947 | #ifndef QT_NO_REGEXP_BACKREF |
2942 | else if ( (yyTok & Tok_BackRef) != 0 ) | 2948 | else if ( (yyTok & Tok_BackRef) != 0 ) |
2943 | box->set( yyTok ^ Tok_BackRef ); | 2949 | box->set( yyTok ^ Tok_BackRef ); |
2944 | #endif | 2950 | #endif |
2945 | else | 2951 | else |
2946 | yyError = TRUE; | 2952 | yyError = TRUE; |
2947 | } | 2953 | } |
2948 | yyTok = getToken(); | 2954 | yyTok = getToken(); |
2949 | } | 2955 | } |
2950 | 2956 | ||
2951 | void QRegExpEngine::parseFactor( Box *box ) | 2957 | void QRegExpEngine::parseFactor( Box *box ) |
2952 | { | 2958 | { |
2953 | #ifndef QT_NO_REGEXP_CAPTURE | 2959 | #ifndef QT_NO_REGEXP_CAPTURE |
2954 | int atom = startAtom( yyMayCapture && yyTok == Tok_LeftParen ); | 2960 | int atom = startAtom( yyMayCapture && yyTok == Tok_LeftParen ); |
2955 | #else | 2961 | #else |
2956 | static const int atom = 0; | 2962 | static const int atom = 0; |
2957 | #endif | 2963 | #endif |
2958 | 2964 | ||
2959 | #ifndef QT_NO_REGEXP_INTERVAL | 2965 | #ifndef QT_NO_REGEXP_INTERVAL |
2960 | #define YYREDO() \ | 2966 | #define YYREDO() \ |
2961 | yyIn = in, yyPos0 = pos0, yyPos = pos, yyLen = len, yyCh = ch, \ | 2967 | yyIn = in, yyPos0 = pos0, yyPos = pos, yyLen = len, yyCh = ch, \ |
2962 | *yyCharClass = charClass, yyMinRep = 0, yyMaxRep = 0, yyTok = tok | 2968 | *yyCharClass = charClass, yyMinRep = 0, yyMaxRep = 0, yyTok = tok |
2963 | 2969 | ||
2964 | const QChar *in = yyIn; | 2970 | const QChar *in = yyIn; |
2965 | int pos0 = yyPos0; | 2971 | int pos0 = yyPos0; |
2966 | int pos = yyPos; | 2972 | int pos = yyPos; |
2967 | int len = yyLen; | 2973 | int len = yyLen; |
2968 | int ch = yyCh; | 2974 | int ch = yyCh; |
2969 | CharClass charClass; | 2975 | CharClass charClass; |
2970 | if ( yyTok == Tok_CharClass ) | 2976 | if ( yyTok == Tok_CharClass ) |
2971 | charClass = *yyCharClass; | 2977 | charClass = *yyCharClass; |
2972 | int tok = yyTok; | 2978 | int tok = yyTok; |
2973 | bool mayCapture = yyMayCapture; | 2979 | bool mayCapture = yyMayCapture; |
2974 | #endif | 2980 | #endif |
2975 | 2981 | ||
2976 | parseAtom( box ); | 2982 | parseAtom( box ); |
2977 | #ifndef QT_NO_REGEXP_CAPTURE | 2983 | #ifndef QT_NO_REGEXP_CAPTURE |
2978 | finishAtom( atom ); | 2984 | finishAtom( atom ); |
2979 | #endif | 2985 | #endif |
2980 | 2986 | ||
2981 | if ( yyTok == Tok_Quantifier ) { | 2987 | if ( yyTok == Tok_Quantifier ) { |
2982 | if ( yyMaxRep == InftyRep ) { | 2988 | if ( yyMaxRep == InftyRep ) { |
2983 | box->plus( atom ); | 2989 | box->plus( atom ); |
2984 | #ifndef QT_NO_REGEXP_INTERVAL | 2990 | #ifndef QT_NO_REGEXP_INTERVAL |
2985 | } else if ( yyMaxRep == 0 ) { | 2991 | } else if ( yyMaxRep == 0 ) { |
2986 | box->clear(); | 2992 | box->clear(); |
2987 | #endif | 2993 | #endif |
2988 | } | 2994 | } |
2989 | if ( yyMinRep == 0 ) | 2995 | if ( yyMinRep == 0 ) |
2990 | box->opt(); | 2996 | box->opt(); |
2991 | 2997 | ||
2992 | #ifndef QT_NO_REGEXP_INTERVAL | 2998 | #ifndef QT_NO_REGEXP_INTERVAL |
2993 | yyMayCapture = FALSE; | 2999 | yyMayCapture = FALSE; |
2994 | int alpha = ( yyMinRep == 0 ) ? 0 : yyMinRep - 1; | 3000 | int alpha = ( yyMinRep == 0 ) ? 0 : yyMinRep - 1; |
2995 | int beta = ( yyMaxRep == InftyRep ) ? 0 : yyMaxRep - ( alpha + 1 ); | 3001 | int beta = ( yyMaxRep == InftyRep ) ? 0 : yyMaxRep - ( alpha + 1 ); |
2996 | 3002 | ||
2997 | Box rightBox( this ); | 3003 | Box rightBox( this ); |
2998 | int i; | 3004 | int i; |
2999 | 3005 | ||
3000 | for ( i = 0; i < beta; i++ ) { | 3006 | for ( i = 0; i < beta; i++ ) { |
3001 | YYREDO(); | 3007 | YYREDO(); |
3002 | Box leftBox( this ); | 3008 | Box leftBox( this ); |
3003 | parseAtom( &leftBox ); | 3009 | parseAtom( &leftBox ); |
3004 | leftBox.cat( rightBox ); | 3010 | leftBox.cat( rightBox ); |
3005 | leftBox.opt(); | 3011 | leftBox.opt(); |
3006 | rightBox = leftBox; | 3012 | rightBox = leftBox; |
3007 | } | 3013 | } |
3008 | for ( i = 0; i < alpha; i++ ) { | 3014 | for ( i = 0; i < alpha; i++ ) { |
3009 | YYREDO(); | 3015 | YYREDO(); |
3010 | Box leftBox( this ); | 3016 | Box leftBox( this ); |
3011 | parseAtom( &leftBox ); | 3017 | parseAtom( &leftBox ); |
3012 | leftBox.cat( rightBox ); | 3018 | leftBox.cat( rightBox ); |
3013 | rightBox = leftBox; | 3019 | rightBox = leftBox; |
3014 | } | 3020 | } |
3015 | rightBox.cat( *box ); | 3021 | rightBox.cat( *box ); |
3016 | *box = rightBox; | 3022 | *box = rightBox; |
3017 | #endif | 3023 | #endif |
3018 | yyTok = getToken(); | 3024 | yyTok = getToken(); |
3019 | #ifndef QT_NO_REGEXP_INTERVAL | 3025 | #ifndef QT_NO_REGEXP_INTERVAL |
3020 | yyMayCapture = mayCapture; | 3026 | yyMayCapture = mayCapture; |
3021 | #endif | 3027 | #endif |
3022 | } | 3028 | } |
3023 | #undef YYREDO | 3029 | #undef YYREDO |
3024 | } | 3030 | } |
3025 | 3031 | ||
3026 | void QRegExpEngine::parseTerm( Box *box ) | 3032 | void QRegExpEngine::parseTerm( Box *box ) |
3027 | { | 3033 | { |
3028 | #ifndef QT_NO_REGEXP_OPTIM | 3034 | #ifndef QT_NO_REGEXP_OPTIM |
3029 | if ( yyTok != Tok_Eos && yyTok != Tok_RightParen && yyTok != Tok_Bar ) | 3035 | if ( yyTok != Tok_Eos && yyTok != Tok_RightParen && yyTok != Tok_Bar ) |
3030 | parseFactor( box ); | 3036 | parseFactor( box ); |
3031 | #endif | 3037 | #endif |
3032 | while ( yyTok != Tok_Eos && yyTok != Tok_RightParen && yyTok != Tok_Bar ) { | 3038 | while ( yyTok != Tok_Eos && yyTok != Tok_RightParen && yyTok != Tok_Bar ) { |
3033 | Box rightBox( this ); | 3039 | Box rightBox( this ); |
3034 | parseFactor( &rightBox ); | 3040 | parseFactor( &rightBox ); |
3035 | box->cat( rightBox ); | 3041 | box->cat( rightBox ); |
3036 | } | 3042 | } |
3037 | } | 3043 | } |
3038 | 3044 | ||
3039 | void QRegExpEngine::parseExpression( Box *box ) | 3045 | void QRegExpEngine::parseExpression( Box *box ) |
3040 | { | 3046 | { |
3041 | parseTerm( box ); | 3047 | parseTerm( box ); |
3042 | while ( yyTok == Tok_Bar ) { | 3048 | while ( yyTok == Tok_Bar ) { |
3043 | Box rightBox( this ); | 3049 | Box rightBox( this ); |
3044 | yyTok = getToken(); | 3050 | yyTok = getToken(); |
3045 | parseTerm( &rightBox ); | 3051 | parseTerm( &rightBox ); |
3046 | box->orx( rightBox ); | 3052 | box->orx( rightBox ); |
3047 | } | 3053 | } |
3048 | } | 3054 | } |
3049 | 3055 | ||
3050 | /* | 3056 | /* |
3051 | The class QRegExpPrivate contains the private data of a regular expression | 3057 | The class QRegExpPrivate contains the private data of a regular expression |
3052 | other than the automaton. It makes it possible for many QRegExp objects to | 3058 | other than the automaton. It makes it possible for many QRegExp objects to |
3053 | use the same QRegExpEngine object with different QRegExpPrivate objects. | 3059 | use the same QRegExpEngine object with different QRegExpPrivate objects. |
3054 | */ | 3060 | */ |
3055 | struct QRegExpPrivate | 3061 | struct QRegExpPrivate |
3056 | { | 3062 | { |
3057 | QString pattern; // regular-expression or wildcard pattern | 3063 | QString pattern; // regular-expression or wildcard pattern |
3058 | QString rxpattern; // regular-expression pattern | 3064 | QString rxpattern; // regular-expression pattern |
3059 | #ifndef QT_NO_REGEXP_WILDCARD | 3065 | #ifndef QT_NO_REGEXP_WILDCARD |
3060 | bool wc; // wildcard mode? | 3066 | bool wc; // wildcard mode? |
3061 | #endif | 3067 | #endif |
3062 | bool min; // minimal matching? (instead of maximal) | 3068 | bool min; // minimal matching? (instead of maximal) |
3063 | #ifndef QT_NO_REGEXP_CAPTURE | 3069 | #ifndef QT_NO_REGEXP_CAPTURE |
3064 | QString t; // last string passed to QRegExp::search() or searchRev() | 3070 | QString t; // last string passed to QRegExp::search() or searchRev() |
3065 | QStringList capturedCache; // what QRegExp::capturedTexts() returned last | 3071 | QStringList capturedCache; // what QRegExp::capturedTexts() returned last |
3066 | #endif | 3072 | #endif |
3067 | QArray<int> captured; // what QRegExpEngine::search() returned last | 3073 | QArray<int> captured; // what QRegExpEngine::search() returned last |
3068 | 3074 | ||
3069 | QRegExpPrivate() { captured.fill( -1, 2 ); } | 3075 | QRegExpPrivate() { captured.fill( -1, 2 ); } |
3070 | }; | 3076 | }; |
3071 | 3077 | ||
3072 | #ifndef QT_NO_REGEXP_OPTIM | 3078 | #ifndef QT_NO_REGEXP_OPTIM |
3073 | static QCache<QRegExpEngine> *engineCache = 0; | 3079 | static QCache<QRegExpEngine> *engineCache = 0; |
3074 | #endif | 3080 | #endif |
3075 | 3081 | ||
3076 | static QRegExpEngine *newEngine( const QString& pattern, bool caseSensitive ) | 3082 | static QRegExpEngine *newEngine( const QString& pattern, bool caseSensitive ) |
3077 | { | 3083 | { |
3078 | #ifndef QT_NO_REGEXP_OPTIM | 3084 | #ifndef QT_NO_REGEXP_OPTIM |
3079 | if ( engineCache != 0 ) { | 3085 | if ( engineCache != 0 ) { |
3080 | QRegExpEngine *eng = engineCache->take( pattern ); | 3086 | QRegExpEngine *eng = engineCache->take( pattern ); |
3081 | if ( eng == 0 || eng->caseSensitive() != caseSensitive ) { | 3087 | if ( eng == 0 || eng->caseSensitive() != caseSensitive ) { |
3082 | delete eng; | 3088 | delete eng; |
3083 | } else { | 3089 | } else { |
3084 | eng->ref(); | 3090 | eng->ref(); |
3085 | return eng; | 3091 | return eng; |
3086 | } | 3092 | } |
3087 | } | 3093 | } |
3088 | #endif | 3094 | #endif |
3089 | return new QRegExpEngine( pattern, caseSensitive ); | 3095 | return new QRegExpEngine( pattern, caseSensitive ); |
3090 | } | 3096 | } |
3091 | 3097 | ||
3092 | static void derefEngine( QRegExpEngine *eng, const QString& pattern ) | 3098 | static void derefEngine( QRegExpEngine *eng, const QString& pattern ) |
3093 | { | 3099 | { |
3094 | if ( eng != 0 && eng->deref() ) { | 3100 | if ( eng != 0 && eng->deref() ) { |
3095 | #ifndef QT_NO_REGEXP_OPTIM | 3101 | #ifndef QT_NO_REGEXP_OPTIM |
3096 | if ( engineCache == 0 ) { | 3102 | if ( engineCache == 0 ) { |
3097 | engineCache = new QCache<QRegExpEngine>; | 3103 | engineCache = new QCache<QRegExpEngine>; |
3098 | engineCache->setAutoDelete( TRUE ); | 3104 | engineCache->setAutoDelete( TRUE ); |
3099 | } | 3105 | } |
3100 | if ( !pattern.isNull() && | 3106 | if ( !pattern.isNull() && |
3101 | engineCache->insert(pattern, eng, 4 + pattern.length() / 4) ) | 3107 | engineCache->insert(pattern, eng, 4 + pattern.length() / 4) ) |
3102 | return; | 3108 | return; |
3103 | #else | 3109 | #else |
3104 | Q_UNUSED( pattern ); | 3110 | Q_UNUSED( pattern ); |
3105 | #endif | 3111 | #endif |
3106 | delete eng; | 3112 | delete eng; |
3107 | } | 3113 | } |
3108 | } | 3114 | } |
3109 | 3115 | ||
3110 | /*! | 3116 | /*! |
3111 | Constructs an empty regexp. | 3117 | Constructs an empty regexp. |
3112 | 3118 | ||
3113 | \sa isValid() | 3119 | \sa isValid() |
3114 | */ | 3120 | */ |
3115 | QRegExp3::QRegExp3() | 3121 | QRegExp3::QRegExp3() |
3116 | { | 3122 | { |
3117 | eng = new QRegExpEngine( TRUE ); | 3123 | eng = new QRegExpEngine( TRUE ); |
3118 | priv = new QRegExpPrivate; | 3124 | priv = new QRegExpPrivate; |
3119 | priv->pattern = QString::null; | 3125 | priv->pattern = QString::null; |
3120 | #ifndef QT_NO_REGEXP_WILDCARD | 3126 | #ifndef QT_NO_REGEXP_WILDCARD |
3121 | priv->wc = FALSE; | 3127 | priv->wc = FALSE; |
3122 | #endif | 3128 | #endif |
3123 | priv->min = FALSE; | 3129 | priv->min = FALSE; |
3124 | compile( TRUE ); | 3130 | compile( TRUE ); |
3125 | } | 3131 | } |
3126 | 3132 | ||
3127 | /*! | 3133 | /*! |
3128 | Constructs a regular expression object for the given \a pattern | 3134 | Constructs a regular expression object for the given \a pattern |
3129 | string. The pattern must be given using wildcard notation if \a | 3135 | string. The pattern must be given using wildcard notation if \a |
3130 | wildcard is TRUE (default is FALSE). The pattern is case sensitive, | 3136 | wildcard is TRUE (default is FALSE). The pattern is case sensitive, |
3131 | unless \a caseSensitive is FALSE. Matching is greedy (maximal), but | 3137 | unless \a caseSensitive is FALSE. Matching is greedy (maximal), but |
3132 | can be changed by calling setMinimal(). | 3138 | can be changed by calling setMinimal(). |
3133 | 3139 | ||
3134 | \sa setPattern() setCaseSensitive() setWildcard() setMinimal() | 3140 | \sa setPattern() setCaseSensitive() setWildcard() setMinimal() |
3135 | */ | 3141 | */ |
3136 | QRegExp3::QRegExp3( const QString& pattern, bool caseSensitive, bool wildcard ) | 3142 | QRegExp3::QRegExp3( const QString& pattern, bool caseSensitive, bool wildcard ) |
3137 | { | 3143 | { |
3138 | eng = 0; | 3144 | eng = 0; |
3139 | priv = new QRegExpPrivate; | 3145 | priv = new QRegExpPrivate; |
3140 | priv->pattern = pattern; | 3146 | priv->pattern = pattern; |
3141 | #ifndef QT_NO_REGEXP_WILDCARD | 3147 | #ifndef QT_NO_REGEXP_WILDCARD |
3142 | priv->wc = wildcard; | 3148 | priv->wc = wildcard; |
3143 | #endif | 3149 | #endif |
3144 | priv->min = FALSE; | 3150 | priv->min = FALSE; |
3145 | compile( caseSensitive ); | 3151 | compile( caseSensitive ); |
3146 | } | 3152 | } |
3147 | 3153 | ||
3148 | /*! | 3154 | /*! |
3149 | Constructs a regular expression as a copy of \a rx. | 3155 | Constructs a regular expression as a copy of \a rx. |
3150 | 3156 | ||
3151 | \sa operator=() | 3157 | \sa operator=() |
3152 | */ | 3158 | */ |
3153 | QRegExp3::QRegExp3( const QRegExp3& rx ) | 3159 | QRegExp3::QRegExp3( const QRegExp3& rx ) |
3154 | { | 3160 | { |
3155 | eng = 0; | 3161 | eng = 0; |
3156 | priv = new QRegExpPrivate; | 3162 | priv = new QRegExpPrivate; |
3157 | operator=( rx ); | 3163 | operator=( rx ); |
3158 | } | 3164 | } |
3159 | 3165 | ||
3160 | /*! | 3166 | /*! |
3161 | Destroys the regular expression and cleans up its internal data. | 3167 | Destroys the regular expression and cleans up its internal data. |
3162 | */ | 3168 | */ |
3163 | QRegExp3::~QRegExp3() | 3169 | QRegExp3::~QRegExp3() |
3164 | { | 3170 | { |
3165 | derefEngine( eng, priv->rxpattern ); | 3171 | derefEngine( eng, priv->rxpattern ); |
3166 | delete priv; | 3172 | delete priv; |
3167 | } | 3173 | } |
3168 | 3174 | ||
3169 | /*! | 3175 | /*! |
3170 | Copies the regular expression \a rx and returns a reference to the copy. | 3176 | Copies the regular expression \a rx and returns a reference to the copy. |
3171 | The case sensitivity, wildcard and minimal matching options are copied as | 3177 | The case sensitivity, wildcard and minimal matching options are copied as |
3172 | well. | 3178 | well. |
3173 | */ | 3179 | */ |
3174 | QRegExp3& QRegExp3::operator=( const QRegExp3& rx ) | 3180 | QRegExp3& QRegExp3::operator=( const QRegExp3& rx ) |
3175 | { | 3181 | { |
3176 | rx.eng->ref(); | 3182 | rx.eng->ref(); |
3177 | derefEngine( eng, priv->rxpattern ); | 3183 | derefEngine( eng, priv->rxpattern ); |
3178 | eng = rx.eng; | 3184 | eng = rx.eng; |
3179 | priv->pattern = rx.priv->pattern; | 3185 | priv->pattern = rx.priv->pattern; |
3180 | priv->rxpattern = rx.priv->rxpattern; | 3186 | priv->rxpattern = rx.priv->rxpattern; |
3181 | #ifndef QT_NO_REGEXP_WILDCARD | 3187 | #ifndef QT_NO_REGEXP_WILDCARD |
3182 | priv->wc = rx.priv->wc; | 3188 | priv->wc = rx.priv->wc; |
3183 | #endif | 3189 | #endif |
3184 | priv->min = rx.priv->min; | 3190 | priv->min = rx.priv->min; |
3185 | #ifndef QT_NO_REGEXP_CAPTURE | 3191 | #ifndef QT_NO_REGEXP_CAPTURE |
3186 | priv->t = rx.priv->t; | 3192 | priv->t = rx.priv->t; |
3187 | priv->capturedCache = rx.priv->capturedCache; | 3193 | priv->capturedCache = rx.priv->capturedCache; |
3188 | #endif | 3194 | #endif |
3189 | priv->captured = rx.priv->captured; | 3195 | priv->captured = rx.priv->captured; |
3190 | return *this; | 3196 | return *this; |
3191 | } | 3197 | } |
3192 | 3198 | ||
3193 | /*! | 3199 | /*! |
3194 | Returns TRUE if this regular expression is equal to \a rx, otherwise | 3200 | Returns TRUE if this regular expression is equal to \a rx, otherwise |
3195 | returns FALSE. | 3201 | returns FALSE. |
3196 | 3202 | ||
3197 | Two QRegExp3 objects are equal if they have the same pattern strings | 3203 | Two QRegExp3 objects are equal if they have the same pattern strings |
3198 | and the same settings for case sensitivity, wildcard and minimal | 3204 | and the same settings for case sensitivity, wildcard and minimal |
3199 | matching. | 3205 | matching. |
3200 | */ | 3206 | */ |
3201 | bool QRegExp3::operator==( const QRegExp3& rx ) const | 3207 | bool QRegExp3::operator==( const QRegExp3& rx ) const |
3202 | { | 3208 | { |
3203 | return priv->pattern == rx.priv->pattern && | 3209 | return priv->pattern == rx.priv->pattern && |
3204 | eng->caseSensitive() == rx.eng->caseSensitive() && | 3210 | eng->caseSensitive() == rx.eng->caseSensitive() && |
3205 | #ifndef QT_NO_REGEXP_WILDCARD | 3211 | #ifndef QT_NO_REGEXP_WILDCARD |
3206 | priv->wc == rx.priv->wc && | 3212 | priv->wc == rx.priv->wc && |
3207 | #endif | 3213 | #endif |
3208 | priv->min == rx.priv->min; | 3214 | priv->min == rx.priv->min; |
3209 | } | 3215 | } |
3210 | 3216 | ||
3211 | /*! \fn bool QRegExp3::operator!=( const QRegExp& rx ) const | 3217 | /*! \fn bool QRegExp3::operator!=( const QRegExp& rx ) const |
3212 | 3218 | ||
3213 | Returns TRUE if this regular expression is not equal to \a rx, otherwise | 3219 | Returns TRUE if this regular expression is not equal to \a rx, otherwise |
3214 | FALSE. | 3220 | FALSE. |
3215 | 3221 | ||
3216 | \sa operator==() | 3222 | \sa operator==() |
3217 | */ | 3223 | */ |
3218 | 3224 | ||
3219 | /*! | 3225 | /*! |
3220 | Returns TRUE if the pattern string is empty, otherwise FALSE. | 3226 | Returns TRUE if the pattern string is empty, otherwise FALSE. |
3221 | 3227 | ||
3222 | If you call match() with an empty pattern on an empty string it will | 3228 | If you call match() with an empty pattern on an empty string it will |
3223 | return TRUE otherwise it returns FALSE since match() operates over the | 3229 | return TRUE otherwise it returns FALSE since match() operates over the |
3224 | whole string. If you call search() with an empty pattern on \e any | 3230 | whole string. If you call search() with an empty pattern on \e any |
3225 | string it will return the start position (0 by default) since it will | 3231 | string it will return the start position (0 by default) since it will |
3226 | match at the start position, because the empty pattern matches the | 3232 | match at the start position, because the empty pattern matches the |
3227 | 'emptiness' at the start of the string, and the length of the match | 3233 | 'emptiness' at the start of the string, and the length of the match |
3228 | returned by matchedLength() will be 0. | 3234 | returned by matchedLength() will be 0. |
3229 | 3235 | ||
3230 | See QString::isEmpty(). | 3236 | See QString::isEmpty(). |
3231 | */ | 3237 | */ |
3232 | 3238 | ||
3233 | bool QRegExp3::isEmpty() const | 3239 | bool QRegExp3::isEmpty() const |
3234 | { | 3240 | { |
3235 | return priv->pattern.isEmpty(); | 3241 | return priv->pattern.isEmpty(); |
3236 | } | 3242 | } |
3237 | 3243 | ||
3238 | /*! | 3244 | /*! |
3239 | Returns TRUE if the regular expression is valid, or FALSE if it's invalid. An | 3245 | Returns TRUE if the regular expression is valid, or FALSE if it's invalid. An |
3240 | invalid regular expression never matches. | 3246 | invalid regular expression never matches. |
3241 | 3247 | ||
3242 | The pattern <b>[a-z</b> is an example of an invalid pattern, since it lacks | 3248 | The pattern <b>[a-z</b> is an example of an invalid pattern, since it lacks |
3243 | a closing square bracket. | 3249 | a closing square bracket. |
3244 | 3250 | ||
3245 | Note that the validity of a regexp may also depend on the setting of | 3251 | Note that the validity of a regexp may also depend on the setting of |
3246 | the wildcard flag, for example <b>*.html</b> is a valid wildcard | 3252 | the wildcard flag, for example <b>*.html</b> is a valid wildcard |
3247 | regexp but an invalid full regexp. | 3253 | regexp but an invalid full regexp. |
3248 | */ | 3254 | */ |
3249 | bool QRegExp3::isValid() const | 3255 | bool QRegExp3::isValid() const |
3250 | { | 3256 | { |
3251 | return eng->isValid(); | 3257 | return eng->isValid(); |
3252 | } | 3258 | } |
3253 | 3259 | ||
3254 | /*! | 3260 | /*! |
3255 | Returns the pattern string of the regular expression. The pattern has either | 3261 | Returns the pattern string of the regular expression. The pattern has either |
3256 | regular expression syntax or wildcard syntax, depending on wildcard(). | 3262 | regular expression syntax or wildcard syntax, depending on wildcard(). |
3257 | 3263 | ||
3258 | \sa setPattern() | 3264 | \sa setPattern() |
3259 | */ | 3265 | */ |
3260 | QString QRegExp3::pattern() const | 3266 | QString QRegExp3::pattern() const |
3261 | { | 3267 | { |
3262 | return priv->pattern; | 3268 | return priv->pattern; |
3263 | } | 3269 | } |
3264 | 3270 | ||
3265 | /*! | 3271 | /*! |
3266 | Sets the pattern string to \a pattern and returns a reference to this regular | 3272 | Sets the pattern string to \a pattern and returns a reference to this regular |
3267 | expression. The case sensitivity, wildcard and minimal matching options are | 3273 | expression. The case sensitivity, wildcard and minimal matching options are |
3268 | not changed. | 3274 | not changed. |
3269 | 3275 | ||
3270 | \sa pattern() | 3276 | \sa pattern() |
3271 | */ | 3277 | */ |
3272 | void QRegExp3::setPattern( const QString& pattern ) | 3278 | void QRegExp3::setPattern( const QString& pattern ) |
3273 | { | 3279 | { |
3274 | if ( priv->pattern != pattern ) { | 3280 | if ( priv->pattern != pattern ) { |
3275 | priv->pattern = pattern; | 3281 | priv->pattern = pattern; |
3276 | compile( caseSensitive() ); | 3282 | compile( caseSensitive() ); |
3277 | } | 3283 | } |
3278 | } | 3284 | } |
3279 | 3285 | ||
3280 | /*! | 3286 | /*! |
3281 | Returns TRUE if case sensitivity is enabled, otherwise FALSE. The default is | 3287 | Returns TRUE if case sensitivity is enabled, otherwise FALSE. The default is |
3282 | TRUE. | 3288 | TRUE. |
3283 | 3289 | ||
3284 | \sa setCaseSensitive() | 3290 | \sa setCaseSensitive() |
3285 | */ | 3291 | */ |
3286 | bool QRegExp3::caseSensitive() const | 3292 | bool QRegExp3::caseSensitive() const |
3287 | { | 3293 | { |
3288 | return eng->caseSensitive(); | 3294 | return eng->caseSensitive(); |
3289 | } | 3295 | } |
3290 | 3296 | ||
3291 | /*! | 3297 | /*! |
3292 | Sets case sensitive matching to \a sensitive. | 3298 | Sets case sensitive matching to \a sensitive. |
3293 | 3299 | ||
3294 | If \a sensitive is TRUE, <b>\\</b><b>.txt$</b> matches | 3300 | If \a sensitive is TRUE, <b>\\</b><b>.txt$</b> matches |
3295 | <tt>readme.txt</tt> but not <tt>README.TXT</tt>. | 3301 | <tt>readme.txt</tt> but not <tt>README.TXT</tt>. |
3296 | 3302 | ||
3297 | \sa caseSensitive() | 3303 | \sa caseSensitive() |
3298 | */ | 3304 | */ |
3299 | void QRegExp3::setCaseSensitive( bool sensitive ) | 3305 | void QRegExp3::setCaseSensitive( bool sensitive ) |
3300 | { | 3306 | { |
3301 | if ( sensitive != eng->caseSensitive() ) | 3307 | if ( sensitive != eng->caseSensitive() ) |
3302 | compile( sensitive ); | 3308 | compile( sensitive ); |
3303 | } | 3309 | } |
3304 | 3310 | ||
3305 | #ifndef QT_NO_REGEXP_WILDCARD | 3311 | #ifndef QT_NO_REGEXP_WILDCARD |
3306 | /*! | 3312 | /*! |
3307 | Returns TRUE if wildcard mode is enabled, otherwise FALSE. The default is | 3313 | Returns TRUE if wildcard mode is enabled, otherwise FALSE. The default is |
3308 | FALSE. | 3314 | FALSE. |
3309 | 3315 | ||
3310 | \sa setWildcard() | 3316 | \sa setWildcard() |
3311 | */ | 3317 | */ |
3312 | bool QRegExp3::wildcard() const | 3318 | bool QRegExp3::wildcard() const |
3313 | { | 3319 | { |
3314 | return priv->wc; | 3320 | return priv->wc; |
3315 | } | 3321 | } |
3316 | 3322 | ||
3317 | /*! Sets the wildcard mode for the regular expression. The default is FALSE. | 3323 | /*! Sets the wildcard mode for the regular expression. The default is FALSE. |
3318 | 3324 | ||
3319 | Setting \a wildcard to TRUE enables simple shell-like wildcard | 3325 | Setting \a wildcard to TRUE enables simple shell-like wildcard |
3320 | matching. | 3326 | matching. |
3321 | (See <a href="#wildcard-matching">wildcard matching (globbing)</a>.) | 3327 | (See <a href="#wildcard-matching">wildcard matching (globbing)</a>.) |
3322 | 3328 | ||
3323 | For example, <b>r*.txt</b> matches the string <tt>readme.txt</tt> in wildcard | 3329 | For example, <b>r*.txt</b> matches the string <tt>readme.txt</tt> in wildcard |
3324 | mode, but does not match <tt>readme</tt>. | 3330 | mode, but does not match <tt>readme</tt>. |
3325 | 3331 | ||
3326 | \sa wildcard() | 3332 | \sa wildcard() |
3327 | */ | 3333 | */ |
3328 | void QRegExp3::setWildcard( bool wildcard ) | 3334 | void QRegExp3::setWildcard( bool wildcard ) |
3329 | { | 3335 | { |
3330 | if ( wildcard != priv->wc ) { | 3336 | if ( wildcard != priv->wc ) { |
3331 | priv->wc = wildcard; | 3337 | priv->wc = wildcard; |
3332 | compile( caseSensitive() ); | 3338 | compile( caseSensitive() ); |
3333 | } | 3339 | } |
3334 | } | 3340 | } |
3335 | #endif | 3341 | #endif |
3336 | 3342 | ||
3337 | /*! Returns TRUE if minimal (non-greedy) matching is enabled, otherwise | 3343 | /*! Returns TRUE if minimal (non-greedy) matching is enabled, otherwise |
3338 | returns FALSE. | 3344 | returns FALSE. |
3339 | 3345 | ||
3340 | \sa setMinimal() | 3346 | \sa setMinimal() |
3341 | */ | 3347 | */ |
3342 | bool QRegExp3::minimal() const | 3348 | bool QRegExp3::minimal() const |
3343 | { | 3349 | { |
3344 | return priv->min; | 3350 | return priv->min; |
3345 | } | 3351 | } |
3346 | 3352 | ||
3347 | /*! | 3353 | /*! |
3348 | Enables or disables minimal matching. If \a minimal is FALSE, matching is | 3354 | Enables or disables minimal matching. If \a minimal is FALSE, matching is |
3349 | greedy (maximal) which is the default. | 3355 | greedy (maximal) which is the default. |
3350 | 3356 | ||
3351 | For example, suppose we have the input string "We must be \<b>bold\</b>, | 3357 | For example, suppose we have the input string "We must be \<b>bold\</b>, |
3352 | very \<b>bold\</b>!" and the pattern <b>\<b>.*\</b></b>. With | 3358 | very \<b>bold\</b>!" and the pattern <b>\<b>.*\</b></b>. With |
3353 | the default greedy (maximal) matching, the match is | 3359 | the default greedy (maximal) matching, the match is |
3354 | "We must be <u>\<b>bold\</b>, very \<b>bold\</b></u>!". But with | 3360 | "We must be <u>\<b>bold\</b>, very \<b>bold\</b></u>!". But with |
3355 | minimal (non-greedy) matching the first match is: | 3361 | minimal (non-greedy) matching the first match is: |
3356 | "We must be <u>\<b>bold\</b></u>, very \<b>bold\</b>!" and the | 3362 | "We must be <u>\<b>bold\</b></u>, very \<b>bold\</b>!" and the |
3357 | second match is | 3363 | second match is |
3358 | "We must be \<b>bold\</b>, very <u>\<b>bold\</b></u>!". | 3364 | "We must be \<b>bold\</b>, very <u>\<b>bold\</b></u>!". |
3359 | In practice we might use the pattern <b>\<b>[^\<]+\</b></b>, | 3365 | In practice we might use the pattern <b>\<b>[^\<]+\</b></b>, |
3360 | although this will still fail for nested tags. | 3366 | although this will still fail for nested tags. |
3361 | 3367 | ||
3362 | \sa minimal() | 3368 | \sa minimal() |
3363 | */ | 3369 | */ |
3364 | void QRegExp3::setMinimal( bool minimal ) | 3370 | void QRegExp3::setMinimal( bool minimal ) |
3365 | { | 3371 | { |
3366 | priv->min = minimal; | 3372 | priv->min = minimal; |
3367 | } | 3373 | } |
3368 | 3374 | ||
3369 | /*! | 3375 | /*! |
3370 | Returns TRUE if \a str is matched exactly by this regular expression | 3376 | Returns TRUE if \a str is matched exactly by this regular expression |
3371 | otherwise it returns FALSE. You can determine how much of the string was | 3377 | otherwise it returns FALSE. You can determine how much of the string was |
3372 | matched by calling matchedLength(). | 3378 | matched by calling matchedLength(). |
3373 | 3379 | ||
3374 | For a given regexp string, R, <tt>match("R")</tt> is the equivalent | 3380 | For a given regexp string, R, <tt>match("R")</tt> is the equivalent |
3375 | of <tt>search("^R$")</tt> since match() effectively encloses the | 3381 | of <tt>search("^R$")</tt> since match() effectively encloses the |
3376 | regexp in the start of string and end of string anchors. | 3382 | regexp in the start of string and end of string anchors. |
3377 | 3383 | ||
3378 | For example, if the regular expression is <b>blue</b>, then match() | 3384 | For example, if the regular expression is <b>blue</b>, then match() |
3379 | returns TRUE only for input <tt>blue</tt>. For inputs | 3385 | returns TRUE only for input <tt>blue</tt>. For inputs |
3380 | <tt>bluebell</tt>, <tt>blutak</tt> and <tt>lightblue</tt>, match() | 3386 | <tt>bluebell</tt>, <tt>blutak</tt> and <tt>lightblue</tt>, match() |
3381 | returns FALSE and matchedLength() will return 4, 3 and 0 respectively. | 3387 | returns FALSE and matchedLength() will return 4, 3 and 0 respectively. |
3382 | 3388 | ||
3383 | \sa search() searchRev() QRegExpValidator | 3389 | \sa search() searchRev() QRegExpValidator |
3384 | */ | 3390 | */ |
3385 | bool QRegExp3::exactMatch( const QString& str ) | 3391 | bool QRegExp3::exactMatch( const QString& str ) |
3386 | { | 3392 | { |
3387 | #ifndef QT_NO_REGEXP_CAPTURE | 3393 | #ifndef QT_NO_REGEXP_CAPTURE |
3388 | priv->t = str; | 3394 | priv->t = str; |
3389 | priv->capturedCache.clear(); | 3395 | priv->capturedCache.clear(); |
3390 | #endif | 3396 | #endif |
3391 | 3397 | ||
3392 | priv->captured = eng->match( str, 0, priv->min, TRUE ); | 3398 | priv->captured = eng->match( str, 0, priv->min, TRUE ); |
3393 | if ( priv->captured[1] == (int) str.length() ) { | 3399 | if ( priv->captured[1] == (int) str.length() ) { |
3394 | return TRUE; | 3400 | return TRUE; |
3395 | } else { | 3401 | } else { |
3396 | priv->captured.detach(); | 3402 | priv->captured.detach(); |
3397 | priv->captured[0] = 0; | 3403 | priv->captured[0] = 0; |
3398 | priv->captured[1] = eng->matchedLength(); | 3404 | priv->captured[1] = eng->matchedLength(); |
3399 | return FALSE; | 3405 | return FALSE; |
3400 | } | 3406 | } |
3401 | } | 3407 | } |
3402 | 3408 | ||
3403 | /*! \overload | 3409 | /*! \overload |
3404 | 3410 | ||
3405 | This version does not set matchedLength(), capturedTexts() and friends. | 3411 | This version does not set matchedLength(), capturedTexts() and friends. |
3406 | */ | 3412 | */ |
3407 | bool QRegExp3::exactMatch( const QString& str ) const | 3413 | bool QRegExp3::exactMatch( const QString& str ) const |
3408 | { | 3414 | { |
3409 | return eng->match(str, 0, priv->min, TRUE)[0] == 0 && | 3415 | return eng->match(str, 0, priv->min, TRUE)[0] == 0 && |
3410 | eng->matchedLength() == (int) str.length(); | 3416 | eng->matchedLength() == (int) str.length(); |
3411 | } | 3417 | } |
3412 | 3418 | ||
3413 | /*! \obsolete | 3419 | /*! \obsolete |
3414 | 3420 | ||
3415 | Attempts to match in \a str, starting from position \a index. Returns the | 3421 | Attempts to match in \a str, starting from position \a index. Returns the |
3416 | position of the match, or -1 if there was no match. | 3422 | position of the match, or -1 if there was no match. |
3417 | 3423 | ||
3418 | The length of the match is stored in \a *len, unless \a len is a null pointer. | 3424 | The length of the match is stored in \a *len, unless \a len is a null pointer. |
3419 | 3425 | ||
3420 | If \a indexIsStart is TRUE (the default), the position \a index in the string | 3426 | If \a indexIsStart is TRUE (the default), the position \a index in the string |
3421 | will match the start of string anchor, <b>^</b>, in the regexp, if present. | 3427 | will match the start of string anchor, <b>^</b>, in the regexp, if present. |
3422 | Otherwise, position 0 in \a str will match. | 3428 | Otherwise, position 0 in \a str will match. |
3423 | 3429 | ||
3424 | Use search() and matchedLength() instead of this function. | 3430 | Use search() and matchedLength() instead of this function. |
3425 | 3431 | ||
3426 | If you really need the \a indexIsStart functionality, try this: | 3432 | If you really need the \a indexIsStart functionality, try this: |
3427 | 3433 | ||
3428 | \code | 3434 | \code |
3429 | QRegExp3 rx( "some pattern" ); | 3435 | QRegExp3 rx( "some pattern" ); |
3430 | int pos = rx.search( str.mid( index ) ); | 3436 | int pos = rx.search( str.mid( index ) ); |
3431 | if ( pos != -1 ) | 3437 | if ( pos != -1 ) |
3432 | pos += index; | 3438 | pos += index; |
3433 | int len = rx.matchedLength(); | 3439 | int len = rx.matchedLength(); |
3434 | \endcode | 3440 | \endcode |
3435 | */ | 3441 | */ |
3436 | #ifndef QT_NO_COMPAT | 3442 | #ifndef QT_NO_COMPAT |
3437 | int QRegExp3::match( const QString& str, int index, int *len, | 3443 | int QRegExp3::match( const QString& str, int index, int *len, |
3438 | bool indexIsStart ) | 3444 | bool indexIsStart ) |
3439 | { | 3445 | { |
3440 | int pos; | 3446 | int pos; |
3441 | if ( indexIsStart ) { | 3447 | if ( indexIsStart ) { |
3442 | pos = search( str.mid(index) ); | 3448 | pos = search( str.mid(index) ); |
3443 | if ( pos >= 0 ) { | 3449 | if ( pos >= 0 ) { |
3444 | pos += index; | 3450 | pos += index; |
3445 | if ( len != 0 ) | 3451 | if ( len != 0 ) |
3446 | *len = matchedLength(); | 3452 | *len = matchedLength(); |
3447 | } else { | 3453 | } else { |
3448 | if ( len != 0 ) | 3454 | if ( len != 0 ) |
3449 | *len = 0; | 3455 | *len = 0; |
3450 | } | 3456 | } |
3451 | } else { | 3457 | } else { |
3452 | pos = search( str, index ); | 3458 | pos = search( str, index ); |
3453 | if ( len != 0 ) | 3459 | if ( len != 0 ) |
3454 | *len = matchedLength(); | 3460 | *len = matchedLength(); |
3455 | } | 3461 | } |
3456 | return pos; | 3462 | return pos; |
3457 | } | 3463 | } |
3458 | #endif | 3464 | #endif |
3459 | 3465 | ||
3460 | /*! | 3466 | /*! |
3461 | Attempts to find a match in \a str from position \a start (0 by default). If | 3467 | Attempts to find a match in \a str from position \a start (0 by default). If |
3462 | \a start is -1, the search starts at the last character; if -2, at the next to | 3468 | \a start is -1, the search starts at the last character; if -2, at the next to |
3463 | last character; etc. | 3469 | last character; etc. |
3464 | 3470 | ||
3465 | Returns the position of the first match, or -1 if there was no match. | 3471 | Returns the position of the first match, or -1 if there was no match. |
3466 | 3472 | ||
3467 | You might prefer to use QString::find(), QString::contains() or even | 3473 | You might prefer to use QString::find(), QString::contains() or even |
3468 | QStringList::grep(). To replace matches use QString::replace(). | 3474 | QStringList::grep(). To replace matches use QString::replace(). |
3469 | 3475 | ||
3470 | Example: | 3476 | Example: |
3471 | \code | 3477 | \code |
3472 | QString str = "offsets: 1.23 .50 71.00 6.00"; | 3478 | QString str = "offsets: 1.23 .50 71.00 6.00"; |
3473 | QRegExp3 rx( "\\d*\\.\\d+" ); // very simple floating point matching | 3479 | QRegExp3 rx( "\\d*\\.\\d+" ); // very simple floating point matching |
3474 | int count = 0; | 3480 | int count = 0; |
3475 | int pos = 0; | 3481 | int pos = 0; |
3476 | while ( pos >= 0 ) { | 3482 | while ( pos >= 0 ) { |
3477 | pos = rx.search( str, pos ); | 3483 | pos = rx.search( str, pos ); |
3478 | count++; | 3484 | count++; |
3479 | } | 3485 | } |
3480 | // pos will be 9, 14, 18 and finally 24; count will end up as 4. | 3486 | // pos will be 9, 14, 18 and finally 24; count will end up as 4. |
3481 | \endcode | 3487 | \endcode |
3482 | 3488 | ||
3483 | \sa searchRev() match() matchedLength() capturedTexts() | 3489 | \sa searchRev() match() matchedLength() capturedTexts() |
3484 | */ | 3490 | */ |
3485 | // QChar versions | 3491 | // QChar versions |
3486 | 3492 | ||
3487 | #ifdef QCHAR_SUPPORT | 3493 | #ifdef QCHAR_SUPPORT |
3488 | const QString makeString(const QChar *str) | 3494 | const QString makeString(const QChar *str) |
3489 | { | 3495 | { |
3490 | // A sentinel value checked in case the QChar *ptr is never null terminated | 3496 | // A sentinel value checked in case the QChar *ptr is never null terminated |
3491 | const uint MAXLENGTH=65535; | 3497 | const uint MAXLENGTH=65535; |
3492 | 3498 | ||
3493 | const QChar *s=str; | 3499 | const QChar *s=str; |
3494 | uint i=0; | 3500 | uint i=0; |
3495 | while(i < MAXLENGTH && *s != QChar::null) { i++;s++ ;} | 3501 | while(i < MAXLENGTH && *s != QChar::null) { i++;s++ ;} |
3496 | return QString(str,i); | 3502 | return QString(str,i); |
3497 | 3503 | ||
3498 | } | 3504 | } |
3499 | int QRegExp3::search(const QChar *str,int start) | 3505 | int QRegExp3::search(const QChar *str,int start) |
3500 | { | 3506 | { |
3501 | return search(makeString(str),start); | 3507 | return search(makeString(str),start); |
3502 | } | 3508 | } |
3503 | int QRegExp3::search(const QChar *str,int start) const | 3509 | int QRegExp3::search(const QChar *str,int start) const |
3504 | { | 3510 | { |
3505 | return search(makeString(str),start); | 3511 | return search(makeString(str),start); |
3506 | } | 3512 | } |
3507 | int QRegExp3::searchRev(const QChar *str,int start) | 3513 | int QRegExp3::searchRev(const QChar *str,int start) |
3508 | { | 3514 | { |
3509 | return searchRev(makeString(str),start); | 3515 | return searchRev(makeString(str),start); |
3510 | } | 3516 | } |
3511 | int QRegExp3::searchRev(const QChar *str,int start) const | 3517 | int QRegExp3::searchRev(const QChar *str,int start) const |
3512 | { | 3518 | { |
3513 | return searchRev(makeString(str),start); | 3519 | return searchRev(makeString(str),start); |
3514 | } | 3520 | } |
3515 | bool QRegExp3::exactMatch(const QChar *str) | 3521 | bool QRegExp3::exactMatch(const QChar *str) |
3516 | { | 3522 | { |
3517 | return exactMatch(makeString(str)); | 3523 | return exactMatch(makeString(str)); |
3518 | } | 3524 | } |
3519 | bool QRegExp3::exactMatch(const QChar *str) const | 3525 | bool QRegExp3::exactMatch(const QChar *str) const |
3520 | { | 3526 | { |
3521 | return exactMatch(makeString(str)); | 3527 | return exactMatch(makeString(str)); |
3522 | } | 3528 | } |
3523 | #endif // QCHAR_SUPPORT | 3529 | #endif // QCHAR_SUPPORT |
3524 | 3530 | ||
3525 | int QRegExp3::search( const QString& str, int start ) | 3531 | int QRegExp3::search( const QString& str, int start ) |
3526 | { | 3532 | { |
3527 | if ( start < 0 ) | 3533 | if ( start < 0 ) |
3528 | start += str.length(); | 3534 | start += str.length(); |
3529 | #ifndef QT_NO_REGEXP_CAPTURE | 3535 | #ifndef QT_NO_REGEXP_CAPTURE |
3530 | priv->t = str; | 3536 | priv->t = str; |
3531 | priv->capturedCache.clear(); | 3537 | priv->capturedCache.clear(); |
3532 | #endif | 3538 | #endif |
3533 | priv->captured = eng->match( str, start, priv->min, FALSE ); | 3539 | priv->captured = eng->match( str, start, priv->min, FALSE ); |
3534 | return priv->captured[0]; | 3540 | return priv->captured[0]; |
3535 | } | 3541 | } |
3536 | 3542 | ||
3537 | /*! \overload | 3543 | /*! \overload |
3538 | 3544 | ||
3539 | This version does not set matchedLength(), capturedTexts() and friends. | 3545 | This version does not set matchedLength(), capturedTexts() and friends. |
3540 | */ | 3546 | */ |
3541 | int QRegExp3::search( const QString& str, int start ) const | 3547 | int QRegExp3::search( const QString& str, int start ) const |
3542 | { | 3548 | { |
3543 | if ( start < 0 ) | 3549 | if ( start < 0 ) |
3544 | start += str.length(); | 3550 | start += str.length(); |
3545 | return eng->match( str, start, priv->min, FALSE )[0]; | 3551 | return eng->match( str, start, priv->min, FALSE )[0]; |
3546 | } | 3552 | } |
3547 | 3553 | ||
3548 | /*! | 3554 | /*! |
3549 | Attempts to find a match backwards in \a str from position \a start. If | 3555 | Attempts to find a match backwards in \a str from position \a start. If |
3550 | \a start is -1 (the default), the search starts at the last character; if -2, | 3556 | \a start is -1 (the default), the search starts at the last character; if -2, |
3551 | at the next to last character; etc. | 3557 | at the next to last character; etc. |
3552 | 3558 | ||
3553 | Returns the position of the first match, or -1 if there was no match. | 3559 | Returns the position of the first match, or -1 if there was no match. |
3554 | 3560 | ||
3555 | You might prefer to use QString::findRev(). | 3561 | You might prefer to use QString::findRev(). |
3556 | 3562 | ||
3557 | \sa search() matchedLength() capturedTexts() | 3563 | \sa search() matchedLength() capturedTexts() |
3558 | */ | 3564 | */ |
3559 | int QRegExp3::searchRev( const QString& str, int start ) | 3565 | int QRegExp3::searchRev( const QString& str, int start ) |
3560 | { | 3566 | { |
3561 | if ( start < 0 ) | 3567 | if ( start < 0 ) |
3562 | start += str.length(); | 3568 | start += str.length(); |
3563 | #ifndef QT_NO_REGEXP_CAPTURE | 3569 | #ifndef QT_NO_REGEXP_CAPTURE |
3564 | priv->t = str; | 3570 | priv->t = str; |
3565 | priv->capturedCache.clear(); | 3571 | priv->capturedCache.clear(); |
3566 | #endif | 3572 | #endif |
3567 | if ( start < 0 || start > (int) str.length() ) { | 3573 | if ( start < 0 || start > (int) str.length() ) { |
3568 | priv->captured.detach(); | 3574 | priv->captured.detach(); |
3569 | priv->captured.fill( -1 ); | 3575 | priv->captured.fill( -1 ); |
3570 | return -1; | 3576 | return -1; |
3571 | } | 3577 | } |
3572 | 3578 | ||
3573 | while ( start >= 0 ) { | 3579 | while ( start >= 0 ) { |
3574 | priv->captured = eng->match( str, start, priv->min, TRUE ); | 3580 | priv->captured = eng->match( str, start, priv->min, TRUE ); |
3575 | if ( priv->captured[0] == start ) | 3581 | if ( priv->captured[0] == start ) |
3576 | return start; | 3582 | return start; |
3577 | start--; | 3583 | start--; |
3578 | } | 3584 | } |
3579 | return -1; | 3585 | return -1; |
3580 | } | 3586 | } |
3581 | 3587 | ||
3582 | /*! \overload | 3588 | /*! \overload |
3583 | 3589 | ||
3584 | This version does not set matchedLength(), capturedText() and friends. | 3590 | This version does not set matchedLength(), capturedText() and friends. |
3585 | */ | 3591 | */ |
3586 | int QRegExp3::searchRev( const QString& str, int start ) const | 3592 | int QRegExp3::searchRev( const QString& str, int start ) const |
3587 | { | 3593 | { |
3588 | if ( start < 0 ) | 3594 | if ( start < 0 ) |
3589 | start += str.length(); | 3595 | start += str.length(); |
3590 | if ( start < 0 || start > (int) str.length() ) | 3596 | if ( start < 0 || start > (int) str.length() ) |
3591 | return -1; | 3597 | return -1; |
3592 | 3598 | ||
3593 | while ( start >= 0 ) { | 3599 | while ( start >= 0 ) { |
3594 | if ( eng->match(str, start, priv->min, TRUE)[0] == start ) | 3600 | if ( eng->match(str, start, priv->min, TRUE)[0] == start ) |
3595 | return start; | 3601 | return start; |
3596 | start--; | 3602 | start--; |
3597 | } | 3603 | } |
3598 | return -1; | 3604 | return -1; |
3599 | } | 3605 | } |
3600 | 3606 | ||
3601 | /*! | 3607 | /*! |
3602 | Returns the length of the last matched string, or -1 if there was no match. | 3608 | Returns the length of the last matched string, or -1 if there was no match. |
3603 | 3609 | ||
3604 | \sa match() search() | 3610 | \sa match() search() |
3605 | */ | 3611 | */ |
3606 | int QRegExp3::matchedLength() | 3612 | int QRegExp3::matchedLength() |
3607 | { | 3613 | { |
3608 | return priv->captured[1]; | 3614 | return priv->captured[1]; |
3609 | } | 3615 | } |
3610 | 3616 | ||
3611 | #ifndef QT_NO_REGEXP_CAPTURE | 3617 | #ifndef QT_NO_REGEXP_CAPTURE |
3612 | /*! | 3618 | /*! |
3613 | Returns a list of the captured text strings. | 3619 | Returns a list of the captured text strings. |
3614 | 3620 | ||
3615 | The first string in the list is the entire matched string. Each | 3621 | The first string in the list is the entire matched string. Each |
3616 | subsequent list element contains a string that matched a | 3622 | subsequent list element contains a string that matched a |
3617 | (capturing) subexpression of the regexp. | 3623 | (capturing) subexpression of the regexp. |
3618 | 3624 | ||
3619 | For example: | 3625 | For example: |
3620 | \code | 3626 | \code |
3621 | QRegExp3 rx( "(\\d+)(\\s*)(cm|inch(es)?)" ); | 3627 | QRegExp3 rx( "(\\d+)(\\s*)(cm|inch(es)?)" ); |
3622 | int pos = rx.search( "Length: 36 inches" ); | 3628 | int pos = rx.search( "Length: 36 inches" ); |
3623 | QStringList list = rx.capturedTexts(); | 3629 | QStringList list = rx.capturedTexts(); |
3624 | // list is now ( "36 inches", "36", " ", "inches", "es" ). | 3630 | // list is now ( "36 inches", "36", " ", "inches", "es" ). |
3625 | \endcode | 3631 | \endcode |
3626 | 3632 | ||
3627 | The above example also captures elements | 3633 | The above example also captures elements |
3628 | that may be present but which we have no interest in. This problem | 3634 | that may be present but which we have no interest in. This problem |
3629 | can be solved by using non-capturing parenthesis: | 3635 | can be solved by using non-capturing parenthesis: |
3630 | 3636 | ||
3631 | \code | 3637 | \code |
3632 | QRegExp3 rx( "(\\d+)(?:\\s*)(cm|inch(?:es)?)" ); | 3638 | QRegExp3 rx( "(\\d+)(?:\\s*)(cm|inch(?:es)?)" ); |
3633 | int pos = rx.search( "Length: 36 inches" ); | 3639 | int pos = rx.search( "Length: 36 inches" ); |
3634 | QStringList list = rx.capturedTexts(); | 3640 | QStringList list = rx.capturedTexts(); |
3635 | // list is now ( "36 inches", "36", "inches" ). | 3641 | // list is now ( "36 inches", "36", "inches" ). |
3636 | \endcode | 3642 | \endcode |
3637 | 3643 | ||
3638 | Some regexps can match an indeterminate number of times. For example | 3644 | Some regexps can match an indeterminate number of times. For example |
3639 | if the input string is "Offsets: 12 14 99 231 7" and the regexp, | 3645 | if the input string is "Offsets: 12 14 99 231 7" and the regexp, |
3640 | <tt>rx</tt>, is <b>(</b><b>\\</b><b>d+)+</b>, we would hope to get a | 3646 | <tt>rx</tt>, is <b>(</b><b>\\</b><b>d+)+</b>, we would hope to get a |
3641 | list of all the numbers matched. However, after calling | 3647 | list of all the numbers matched. However, after calling |
3642 | <tt>rx.search(str)</tt>, capturedTexts() will return the list ( "12", | 3648 | <tt>rx.search(str)</tt>, capturedTexts() will return the list ( "12", |
3643 | "12" ), i.e. the entire match was "12" and the first subexpression | 3649 | "12" ), i.e. the entire match was "12" and the first subexpression |
3644 | matched was "12". The correct approach is to use cap() in a | 3650 | matched was "12". The correct approach is to use cap() in a |
3645 | <a href="#cap_in_a_loop">loop</a>. | 3651 | <a href="#cap_in_a_loop">loop</a>. |
3646 | 3652 | ||
3647 | The order of elements in the string list is as follows. The first | 3653 | The order of elements in the string list is as follows. The first |
3648 | element is the entire matching string. Each subsequent element | 3654 | element is the entire matching string. Each subsequent element |
3649 | corresponds to the next capturing open left parenthesis. Thus | 3655 | corresponds to the next capturing open left parenthesis. Thus |
3650 | capturedTexts()[1] is the text of the first capturing parenthesis, | 3656 | capturedTexts()[1] is the text of the first capturing parenthesis, |
3651 | capturedTexts()[2] is the text of the second and so on (corresponding | 3657 | capturedTexts()[2] is the text of the second and so on (corresponding |
3652 | to $1, $2 etc. in some other regexp languages). | 3658 | to $1, $2 etc. in some other regexp languages). |
3653 | 3659 | ||
3654 | \sa cap() pos() | 3660 | \sa cap() pos() |
3655 | */ | 3661 | */ |
3656 | QStringList QRegExp3::capturedTexts() | 3662 | QStringList QRegExp3::capturedTexts() |
3657 | { | 3663 | { |
3658 | if ( priv->capturedCache.isEmpty() ) { | 3664 | if ( priv->capturedCache.isEmpty() ) { |
3659 | for ( int i = 0; i < (int) priv->captured.size(); i += 2 ) { | 3665 | for ( int i = 0; i < (int) priv->captured.size(); i += 2 ) { |
3660 | QString m; | 3666 | QString m; |
3661 | if ( priv->captured[i + 1] == 0 ) | 3667 | if ( priv->captured[i + 1] == 0 ) |
3662 | m = QString::fromLatin1( "" ); | 3668 | m = QString::fromLatin1( "" ); |
3663 | else if ( priv->captured[i] >= 0 ) | 3669 | else if ( priv->captured[i] >= 0 ) |
3664 | m = priv->t.mid( priv->captured[i], | 3670 | m = priv->t.mid( priv->captured[i], |
3665 | priv->captured[i + 1] ); | 3671 | priv->captured[i + 1] ); |
3666 | priv->capturedCache.append( m ); | 3672 | priv->capturedCache.append( m ); |
3667 | } | 3673 | } |
3668 | priv->t = QString::null; | 3674 | priv->t = QString::null; |
3669 | } | 3675 | } |
3670 | return priv->capturedCache; | 3676 | return priv->capturedCache; |
3671 | } | 3677 | } |
3672 | 3678 | ||
3673 | /*! Returns the text captured by the \a nth subexpression. The entire match | 3679 | /*! Returns the text captured by the \a nth subexpression. The entire match |
3674 | has index 0 and the parenthesised subexpressions have indices starting | 3680 | has index 0 and the parenthesised subexpressions have indices starting |
3675 | from 1 (excluding non-capturing parenthesis). | 3681 | from 1 (excluding non-capturing parenthesis). |
3676 | 3682 | ||
3677 | \code | 3683 | \code |
3678 | QRegExp3 rxlen( "(\\d+)(?:\\s*)(cm|inch)" ); | 3684 | QRegExp3 rxlen( "(\\d+)(?:\\s*)(cm|inch)" ); |
3679 | int pos = rxlen.search( "Length: 189cm" ); | 3685 | int pos = rxlen.search( "Length: 189cm" ); |
3680 | if ( pos > -1 ) { | 3686 | if ( pos > -1 ) { |
3681 | QString value = rxlen.cap( 1 );// "189" | 3687 | QString value = rxlen.cap( 1 ); // "189" |
3682 | QString unit = rxlen.cap( 2 ); // "cm" | 3688 | QString unit = rxlen.cap( 2 ); // "cm" |
3683 | // ... | 3689 | // ... |
3684 | } | 3690 | } |
3685 | \endcode | 3691 | \endcode |
3686 | 3692 | ||
3687 | <a name="cap_in_a_loop"> | 3693 | <a name="cap_in_a_loop"> |
3688 | Some patterns may lead to a number of matches which cannot be | 3694 | Some patterns may lead to a number of matches which cannot be |
3689 | determined in advance, for example:</a> | 3695 | determined in advance, for example:</a> |
3690 | 3696 | ||
3691 | \code | 3697 | \code |
3692 | QRegExp3 rx( "(\\d+)" ); | 3698 | QRegExp3 rx( "(\\d+)" ); |
3693 | str = "Offsets: 12 14 99 231 7"; | 3699 | str = "Offsets: 12 14 99 231 7"; |
3694 | QStringList list; | 3700 | QStringList list; |
3695 | pos = 0; | 3701 | pos = 0; |
3696 | while ( pos >= 0 ) { | 3702 | while ( pos >= 0 ) { |
3697 | pos = rx.search( str, pos ); | 3703 | pos = rx.search( str, pos ); |
3698 | if ( pos > -1 ) { | 3704 | if ( pos > -1 ) { |
3699 | list += rx.cap( 1 ); | 3705 | list += rx.cap( 1 ); |
3700 | pos += rx.matchedLength(); | 3706 | pos += rx.matchedLength(); |
3701 | } | 3707 | } |
3702 | } | 3708 | } |
3703 | // list contains: ( "12", "14", "99", "231", "7" ). | 3709 | // list contains: ( "12", "14", "99", "231", "7" ). |
3704 | \endcode | 3710 | \endcode |
3705 | 3711 | ||
3706 | The order of elements matched by cap() is as follows. The first | 3712 | The order of elements matched by cap() is as follows. The first |
3707 | element, cap( 0 ), is the entire matching string. Each subsequent | 3713 | element, cap( 0 ), is the entire matching string. Each subsequent |
3708 | element corresponds to the next capturing open left parenthesis. Thus | 3714 | element corresponds to the next capturing open left parenthesis. Thus |
3709 | cap( 1 ) is the text of the first capturing parenthesis, cap( 2 ) is | 3715 | cap( 1 ) is the text of the first capturing parenthesis, cap( 2 ) is |
3710 | the text of the second and so on. | 3716 | the text of the second and so on. |
3711 | 3717 | ||
3712 | \sa search() pos() capturedTexts() | 3718 | \sa search() pos() capturedTexts() |
3713 | */ | 3719 | */ |
3714 | QString QRegExp3::cap( int nth ) | 3720 | QString QRegExp3::cap( int nth ) |
3715 | { | 3721 | { |
3716 | if ( nth < 0 || nth >= (int) priv->captured.size() / 2 ) | 3722 | if ( nth < 0 || nth >= (int) priv->captured.size() / 2 ) |
3717 | return QString::null; | 3723 | return QString::null; |
3718 | else | 3724 | else |
3719 | return capturedTexts()[nth]; | 3725 | return capturedTexts()[nth]; |
3720 | } | 3726 | } |
3721 | 3727 | ||
3722 | /*! Returns the position of the \a nth captured text in the | 3728 | /*! Returns the position of the \a nth captured text in the |
3723 | searched string. If \a nth is 0 (the default), pos() returns the | 3729 | searched string. If \a nth is 0 (the default), pos() returns the |
3724 | position of the whole match. | 3730 | position of the whole match. |
3725 | 3731 | ||
3726 | Example: | 3732 | Example: |
3727 | \code | 3733 | \code |
3728 | QRegExp3 rx( "/([a-z]+)/([a-z]+)" ); | 3734 | QRegExp3 rx( "/([a-z]+)/([a-z]+)" ); |
3729 | rx.search( "Output /dev/null" );// Returns 7 (position of /dev/null) | 3735 | rx.search( "Output /dev/null" ); // Returns 7 (position of /dev/null) |
3730 | rx.pos( 0 ); // Returns 7 (position of /dev/null) | 3736 | rx.pos( 0 ); // Returns 7 (position of /dev/null) |
3731 | rx.pos( 1 ); // Returns 8 (position of dev) | 3737 | rx.pos( 1 ); // Returns 8 (position of dev) |
3732 | rx.pos( 2 ); // Returns 12 (position of null) | 3738 | rx.pos( 2 ); // Returns 12 (position of null) |
3733 | \endcode | 3739 | \endcode |
3734 | 3740 | ||
3735 | Note that pos() returns -1 for zero-length matches. (For example, if | 3741 | Note that pos() returns -1 for zero-length matches. (For example, if |
3736 | cap(4) would return an empty string, pos(4) returns -1.) This is due | 3742 | cap(4) would return an empty string, pos(4) returns -1.) This is due |
3737 | to an implementation tradeoff. | 3743 | to an implementation tradeoff. |
3738 | 3744 | ||
3739 | \sa capturedTexts() cap() | 3745 | \sa capturedTexts() cap() |
3740 | */ | 3746 | */ |
3741 | int QRegExp3::pos( int nth ) | 3747 | int QRegExp3::pos( int nth ) |
3742 | { | 3748 | { |
3743 | if ( nth < 0 || nth >= (int) priv->captured.size() / 2 ) | 3749 | if ( nth < 0 || nth >= (int) priv->captured.size() / 2 ) |
3744 | return -1; | 3750 | return -1; |
3745 | else | 3751 | else |
3746 | return priv->captured[2 * nth]; | 3752 | return priv->captured[2 * nth]; |
3747 | } | 3753 | } |
3748 | #endif | 3754 | #endif |
3749 | 3755 | ||
3750 | void QRegExp3::compile( bool caseSensitive ) | 3756 | void QRegExp3::compile( bool caseSensitive ) |
3751 | { | 3757 | { |
3752 | derefEngine( eng, priv->rxpattern ); | 3758 | derefEngine( eng, priv->rxpattern ); |
3753 | #ifndef QT_NO_REGEXP_WILDCARD | 3759 | #ifndef QT_NO_REGEXP_WILDCARD |
3754 | if ( priv->wc ) | 3760 | if ( priv->wc ) |
3755 | priv->rxpattern = wc2rx( priv->pattern ); | 3761 | priv->rxpattern = wc2rx( priv->pattern ); |
3756 | else | 3762 | else |
3757 | #endif | 3763 | #endif |
3758 | priv->rxpattern = priv->pattern.isNull() ? QString::fromLatin1( "" ) | 3764 | priv->rxpattern = priv->pattern.isNull() ? QString::fromLatin1( "" ) |
3759 | : priv->pattern; | 3765 | : priv->pattern; |
3760 | eng = newEngine( priv->rxpattern, caseSensitive ); | 3766 | eng = newEngine( priv->rxpattern, caseSensitive ); |
3761 | #ifndef QT_NO_REGEXP_CAPTURE | 3767 | #ifndef QT_NO_REGEXP_CAPTURE |
3762 | priv->t = QString::null; | 3768 | priv->t = QString::null; |
3763 | priv->capturedCache.clear(); | 3769 | priv->capturedCache.clear(); |
3764 | #endif | 3770 | #endif |
3765 | priv->captured.detach(); | 3771 | priv->captured.detach(); |
3766 | priv->captured.fill( -1, 2 + 2 * eng->numCaptures() ); | 3772 | priv->captured.fill( -1, 2 + 2 * eng->numCaptures() ); |
3767 | } | 3773 | } |