robustness improvements in link descovery mechanisms

accept stray spaces in link tags and inside relevant attribute values.

Thanks, Gen, for pointing it out.

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

1 files changed, 4 insertions, 4 deletions

diff --git a/lib/consumer.cc b/lib/consumer.cc
index 299b3bc..ff5da91 100644
--- a/lib/consumer.cc
+++ b/lib/consumer.cc
@@ -278,28 +278,28 @@ namespace opkele {
 	CURLcode r;
 	(r=curl_misc_sets(curl))
 	|| (r=curl_easy_setopt(curl,CURLOPT_URL,url.c_str()))
 	|| (r=curl_easy_setopt(curl,CURLOPT_WRITEFUNCTION,_curl_tostring))
 	|| (r=curl_easy_setopt(curl,CURLOPT_WRITEDATA,&html))
 	;
 	if(r)
 	    throw exception_curl(OPKELE_CP_ "failed to curl_easy_setopt()",r);
 	r = curl_easy_perform(curl);
 	if(r && r!=CURLE_WRITE_ERROR)
 	    throw exception_curl(OPKELE_CP_ "failed to curl_easy_perform()",r);
 	// strip out everything past body
-	static const char *re_hdre = "<head[^>]*>",
-		     *re_lre = "<link\\b([^>]+)>",
-		     *re_rre = "\\brel=['\"]([^'\"]+)['\"]",
-		     *re_hre = "\\bhref=['\"]([^'\"]+)['\"]";
+	static const char *re_hdre = "<\\s*head[^>]*>",
+		     *re_lre = "<\\s*link\\b([^>]+)>",
+		     *re_rre = "\\brel\\s*=\\s*['\"]\\s*([^'\"\\s]+)\\s*['\"]",
+		     *re_hre = "\\bhref\\s*=\\s*['\"]\\s*([^'\"\\s]+)\\s*['\"]";
 #if defined(USE_LIBPCRECPP)
 	static pcrecpp::RE_Options ro(PCRE_CASELESS|PCRE_DOTALL);
 	static pcrecpp::RE
 	    bre("<body\\b.*",ro), hdre(re_hdre,ro),
 	    lre(re_lre,ro), rre(re_rre), hre(re_hre,ro);
 	bre.Replace("",&html);
 	pcrecpp::StringPiece hpiece(html);
 	if(!hdre.FindAndConsume(&hpiece))
 	    throw bad_input(OPKELE_CP_ "failed to find head");
 	string attrs;
 	while(lre.FindAndConsume(&hpiece,&attrs)) {
 	    pcrecpp::StringPiece rel, href;