Re: lynx-dev 2.8.4dev.3 ALIGN=center bugs

[Klaus Weide]

On Fri, 9 Jun 2000, Andrey A. Chernov wrote:

> Consider to view HTML file below and you notice strange things which means
> that:

This isn't new in 2.8.3 in 2.8.4dev, it's been around for a long time.

> 1) <TR ALIGN=center> does screen-center instead of centering in the table,
> but it is lesser bug (lack of tables implementation feature?)

(yes...)

> 2) It is greater bug: ALIGN=center not turned off since first occurance,
> so next TR (without ALIGN) appearse centered too in the same manner as 1)

I have seen strange things like that, but never in a simple example
like yours.

The way alignment inheritance is handled in HTML.c is, uhm, confusing
and inconsistent.  One day it all ought to be cleaned up...

Here is a band-aid fix, works only in sorta SGML mode (no -tagsoup).
Insert in HTML_end_element (src/HTML.c), under

    case HTML_TR:

but before

        me->in_word = NO;

----------------- snip -------------------
        if (LYoverride_default_alignment(me)) {
            ;                   /* nothing */
        } else if (me->List_Nesting_Level >= 0 ||
                   ((me->Division_Level < 0) &&
                    (!strcmp(me->sp->style->name, "Normal") ||
                     !strcmp(me->sp->style->name, "Preformatted")))) {
            ;                   /* nothing */
        } else {
            if (me->sp > me->stack &&
                me->sp[0].style == me->sp[-1].style &&
                me->Division_Level >= 0 &&
                (int)me->DivisionAlignments[me->Division_Level]
                == me->current_default_alignment)
                me->sp[0].style->alignment = (short) 
me->current_default_alignment;
        }
----------------- snip -------------------

Not tested with original HTML.c, only with one that had some
other changes.  The checks are probably overkill, but just in
case.  There may be similar cases where this doesn't help.


  Klaus


; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden