[links-list] Re: probable bug in elinks: wrong < and > interpretation

Khimenko Victor khim at sch57.msk.ru
Fri Jan 4 07:29:01 PST 2002



On Fri, 4 Jan 2002, Petr Baudis wrote:

> Dear diary, on Fri, Jan 04, 2002 at 07:16:07AM CET, I got a letter, where Anton
> Voloshin <vav at isv.ru> told me, that...
> > using current development elinks to look at
> > http://www.research.att.com/~bs/compilers.html, I've noticed strange looking
> > line in program example:
> >
> > cout <> s;
> >
> > E.g. galeon shows the same place more correctly:
> >
> > cout << "Please enter your first name followed by a newline\n"; cin >> s;
> >
> > I'm not sure that this is a bug (strange usage of < and > symbols in HTML),
> > but I'd like links to show such pages in a more nice way anyway. :)
> Well, it IS start and end of a tag, and elinks interpreted it correctly.  IMHO
> it would maybe be possible to hack some kind of more fuzzy logic to HTML
> parser, but then we would lose our pedantic aura ;-)). And we wouldn't conform
> to standarts so much as we do now. I just don't happen to like this a lot.
>
> The solution which comes to mind are two modes of HTML parsing - first one
> strict, and the second one relaxed, which would accept this. However I've no
> idea if someone would use the strict mode then after all ;-).
>

IMO better not strict vs relaxed but more like "try to normalize broken
page" mode. Some additional parser which can be called from menu or via
hotkey when page is fully loaded and which tries to make page more
standard-compilant. Since page is fully in memory it can use more
versions of fuzzy logic. I'm not sure if it worth the trouble, though:
"normal pages" can be viewed already and no amount of fuzzy logic will
make javascript-based or flash-based site working in links anyway...

-- 
Unsubscribe: send email to links-list-request at linuxfromscratch.org
and put unsubscribe in the subject header of the message



More information about the links-list mailing list