[Htmlparser-announce] htmlparser 1.42

[Derrick O. <Der...@Ro...>]

Patch release Version 1.42 (Release Build Jul 27, 2004) of the most 
popular HTML parser on SourceForge is now available:

http://sourceforge.net/project/showfiles.php?group_id=24399&package_id=17243&release_id=256305

This is the same as Version 1.4 with four bug fixes:

#998195 SiteCatpurer just crashed
#995744 Translate.decode(String)
#995703 Parser Crash
#988846 Linkbean getLinks() segmentation fault (duplicate of above)
#919738 Text has not been extracted correctly using StringBean
#936392 ScriptTag visitor fails for comments with ' (duplicate of above)


One bug involved the decoding of URLs with the Translate.decode() 
method, which was incorrect.

Another bug involved the SiteCapturer program failing in the face of an 
EncodingChangeException. This exception is raised when the <META> tag 
indicates a different character set that the one assumed at the start of 
parsing, and retracing the stream yields different characters than those 
the client has already consumed. The SiteCapturer now handles this 
exception by resetting the parser and trying again.

Another bug involved an overzealous test for "text/XXX" content. This 
was erroneously assuming that the content was binary and throwing an 
exception. Experience indicates numerous web servers are returning 
parseable streams with content types that do not indicate text. The test 
is removed.

The last bug would return wrong nodes when presented with a quote in a 
comment in a <SCRIPT> tag.