Re: Bugreport on comments
hellmuth@ira.uka.de
Fri, 5 Dec 1997 00:17:53 +0100 (MET)
Ok, my fault. I saw the netscape_buggy_comment switch, but thought it was
addressing a different problem, because of the follwing comment in Parser.pm:
#
# Netscape does not parse this comment correctly (it terminates the comment
# too early):
#
# <! -- comment -- --> more comment -->
#
I have just looked up the HTML2.0 definition of comments, which seems to have
changed in HTML4.0 to disallow multiple comments between markup delimiters
(step in the right direction IMHO, comments shouldn't be syntactical beasts
that break silently when you make a mistake). So, when 4.0 will be out, you
are in a dilemma anyway: If you try to be 4.0 compliant, programms using libwww
can't parse older texts correctly (or is something wrong with my
reasoning here). As it stands now, libwww can't parse a lot of texts floating
on the web, and without telling anyone (how about giving a warning message,
UNLESS netscape_buggy_comment is true?).
It's no wonder that there are a lot of SGML-challenged users when for example
O'Reilly's 'HTML: THe definitve Guide' forgets to mention the facts about
comment use. Suns HotJava browser ignores the standard too, not suprisingly
since I stumbled over the <--------> comments while loading one of their
Web-pages. How about renaming netscape_buggy_comment to buggy_comment.
Holger.