updated news and bumped version to 2.0.32.0.3public/master

Signed-off-by: Michael Krelin <hacker@klever.net>

1 files changed, 5 insertions, 0 deletions

diff --git a/NEWS.xml b/NEWS.xml
index 0206018..93d88f9 100644
--- a/NEWS.xml
+++ b/NEWS.xml
@@ -1,66 +1,71 @@
 <?xml version="1.0" encoding="us-ascii"?>
 <news>
+ <version version="2.0.3" date="September 1st, 2009">
+  <ni>Fixed an URI normalization bug misnormalizing URIs with a single-character
+  trailing component</ni>
+  <ni>build improvements</ni>
+ </version>
  <version version="2.0.2" date="April 11th, 2009">
   <ni>Handling of unknown encodings during discovery</ni>
   <ni>Discovery robustness improvements</ni>
   <ni>Workaround for OPs (e.g. livejournal.com) breaking specs</ni>
   <ni>Build fixes and improvements</ni>
  </version>
  <version version="2.0.1" date="November 22nd, 2008">
   <ni>Compile-time fixes and improvements</ni>
   <ni>Portability improvements for FreeBSD</ni>
   <ni>Really suppress debugging message from htmltidy when --disable-debug is in
   effect</ni>
   <ni>minor bugfixes</ni>
   <ni>thread-safety improvements</ni>
  </version>
  <version version="2.0" date="June 26th, 2008">
   <ni>OpenID 2.0 support</ni>
   <ni>Major rewrite of the whole thing</ni>
   <ni>Support for XRDS (YADIS and XRI/inames) discovery</ni>
   <ni>Sheerly improved html-based discovery (only code using new, 2.0-enabled
   classes benefits from it)</ni>
   <ni>Deprecation of the old api</ni>
   <ni>Added sample RP and OP implementations</ni>
   <ni>Require expat xml stream parser library</ni>
   <ni>Require htmltidy library</ni>
   <ni>Require tr1/memory (shared_ptr) support - either modern gcc or boost
   library</ni>
  </version>
  <version version="0.3.2" date="November 22nd, 2007">
   <ni>code cleanup for stricter compiler</ni>
  </version>
  <version version="0.3.1" date="November 20th, 2007">
   <ni>more robustness improvements in links discovery</ni>
   <ni>removed dependency on pcre c++ bindings, because there are few of them and
   not everyone likes all of them</ni>
   <ni>minor build improvements</ni>
  </version>
  <version version="0.3" date="August 9th, 2007">
   <ni>fixed canonicalization procedure to be specs-compliant. Note, that the old
   consumer_t::canonicalize is now called consumer_t::normalize and the
   canonicalize memeber is now virtual to allow caching layer, not static</ni>
   <ni>robustness improvement in handling associations expiry</ni>
   <ni>minor documentation updates</ni>
  </version>
  <version version="0.2.1" date="June 24th, 2007">
   <ni>open id server invalid signature bugfix</ni>
  </version>
  <version version="0.2" date="June 19th, 2007">
   <ni>A few robustness improvements and optimizations</ni>
   <ni>More liberal key/values messages parsing</ni>
   <ni>Changed unusable --with-pcre++ configure option to --with-pcrepp</ni>
  </version>
  <version version="0.1.1" date="January 16th, 2007">
   <ni>Fixed a bug in curl errors handling</ni>
   <ni>added --disable-ssl-verify-host and --disable-ssl-verify-peer options to
   configure to alter default implementation of consumer's retrieve_links
   behaviour</ni>
   <ni>Build fix that eliminates the need to pass --disable-konforka in the
   absence of it.</ni>
  </version>
  <version version="0.1" date="January 12th, 2007">
   <ni>OpenID simple registration extension implementation</ni>
   <ni>OpenID extensions framework</ni>
   <ni>Canonicalization bugfix</ni>
   <ni>Slightly improved interoperability with buggy implementations</ni>