Mailing List Archive

> I've uploaded B59-include-bugs-needs-B46.txt which fixes several bugs
> in http_include.c's parsing routines:
> Bugs fixed: (decreasing importance)

> * did not allow whitespace around = in tag=value, or between -- and >

> David.


Without checking the HTML spec, isn't a comment supposed to end
with --> with now white space ? .... we might be encouraging sloppy
HTML.

n.b. a few browsers for NeXt barf unless they see a --> terminated
comment - they just stay in comment mode, so big chunks of documents
disappear. While includes are a special case of comment, it might
be better to be strict here too.


robh

> I've uploaded B59-include-bugs-needs-B46.txt which fixes several bugs
> in http_include.c's parsing routines:
> Bugs fixed: (decreasing importance)
> * Couldn't have \ in tag value strings.
> * Couldn't have html entities in tag values, such as "

> * #echo did not html encode & < > in the string it echoed

Er, do we want to be making the server rewrite ANY text that is in the
control of users? If the solution is for the user to write &lt; &gt;
in their #echo then it should be up to them.

I'm probably missing the point, but as a general rule I think the server
shouldn't modify any text the user supplies. If the user wants to break
her/his browser then that's their lookout.

> * #directive names were case sensitive

Er, what does that mean? Got an example?

> * tag names were case sensitive
> * tag value strings should be delimited by " _or_ '
> * did not allow whitespace around = in tag=value, or between -- and >


Right, I'd seen that, never thought of it as a bug before tho. Good.

> David.
>

Ay.

Andrew Wilson URL: http://www.cm.cf.ac.uk/User/Andrew.Wilson/
Elsevier Science, Oxford Office: +44 01865 843155 Mobile: +44 0589 616144

> Without checking the HTML spec, isn't a comment supposed to end
> with --> with now white space ? .... we might be encouraging sloppy
> HTML.

The only safe comment is exactly

<!-- comment (not including "--" or ">") -->

but space is allowed between the "--" and ">".

True SGML comments are a bit more complicated.
See <http://www.acl.lanl.gov/HTML_WG/html-wg-95q1.messages/0189.html>

.......Roy