[Htmlparser-developer] General Bug Behavior

SourceForge Headquarters 1320 Columbia Street Suite 310 San Diego, CA 92101 +1 (858) 422-6466

We've found three documents over the last few days that cause the
HTMLParser to hang. I will make sure they get into the bug database but
the issue centers around what should happen when the parser encounters
ill-formed HTML. I would propose that the correct behavior is to throw
and exception if the parser is unable to handle the syntax, but right
now it just hangs. Clearly, more investigation is required to determine
whether it's in a loop or waiting on the input. Since I'm not sure what
a fix would entail, I though it worth raising the issue as a general
design question. What should be done when the parser encounters
malformed HTML that goes beyond the realm of reasonable recovery?
=20
BTW: The documents we encountered that hung the parser had the following
artifacts:
=20
1) Inclusiong of "<!-->" pattern which is technically an invalid comment
syntax.
2) Inclusion of the "<html><head><TITLE>" pattern twice at the beginning
of the document.
3) Two opening "<TITLE>" tags with only one ending "</TITLE>" tag.
=20
From our point of view, a hag is devastating in that it does not allow
the application to move forward. An exception would be ideal in that it
would identify the problem without breaking the application.
=20