Moin moin,
as announced by mirabile, here my observation concerning prettysrc.
(we're sitting on [EMAIL PROTECTED] for those who care ;)
If you have this example HTML source
------ QUOTE BEGIN ------
<html> <body>
<p nowrap width="65%">
blub
</p>
</body> </html>
------- QUOTE END -------
and have
PRETTYSRC_SPEC:ATTRIB:b:!b
in your lynx.cfg, then the "nowrap" will cause the rest of the
file to be bold, ignoring all other tags and their prettysrc
specs.
If you have such a "nowrap" at the end of a tag, then the rest
looks fine (like <p width="65%" nowrap>).
If I understand WWW/Library/Implementation/SGML.c right, then this
happens because of the various
------ QUOTE BEGIN ------
if (WHITE(c))
break;
------- QUOTE END -------
parts for several cases of "switch (context->state) {",
like S_attr, S_attr_gap, S_equals, S_value and possibly others
where a WHITE breaks an element but the prettyprinting shouldn't
go nuts because of that.
At each occurance of check for '>' lynx should extend the check
for WHITE to let it execute PSRCSTOP when WHITE happens.
Even though an attribute without value assignment might not be
correct by standard, it still shouldn't break the prettyprinting.
Or is there are reasoning for breaking the prettyprinting?
Why not treat prematurely WHITE-ended attribute as "badattr" instead?
(beware of my reasoning, I'm just a user, no technical expert,
neither for standards {i.e. what should be correct} nor lynx code
{whether it is correctly implemented})
--
© Rado S. -- You must provide YOUR effort for your goal!
Even if it seems insignificant, in fact EVERY effort counts
for a shared task, at least to show your deserving attitude.
_______________________________________________
Lynx-dev mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/lynx-dev
