[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: lynx-dev Re: [PATCH 2.8.5-dev15] table horizontal shift
From: |
Thomas Dickey |
Subject: |
Re: lynx-dev Re: [PATCH 2.8.5-dev15] table horizontal shift |
Date: |
Mon, 6 Oct 2003 21:04:47 -0400 |
User-agent: |
Mutt/1.3.27i |
On Mon, Oct 06, 2003 at 09:36:01PM -0300, Frdric L. W. Meunier wrote:
> On Mon, 6 Oct 2003, Thomas Dickey wrote:
>
> > On Mon, 6 Oct 2003, [ISO-8859-1] Fr?d?ric L. W. Meunier wrote:
> >
> > > With these last 2 it no longer crashes with the links Thomas
> > > and I reported, but it does at
> > > http://www.google.com/search?q=lynx&btnG=Search when I disable
> > > nested tables with "~".
> >
> > I didn't see that, but had in mind that I should look for things like
> > that. I browsed a little on groups.google.com, noticed that some
> > parts of the table layout are improved - but others are worse (it's
> > hard to say what the total effect is).
>
> Does that mean you don't get it to crash with your 2 patches +
> the 2 from Ilya disabling nested tables ? But here's what I
yes - I was running color-styles and as you suggested disabled nested-tables.
Just in case it was something that valgrind could find, I also tried with
that - neither broke. (Sometimes running valgrind keeps the application
running, so I tried it both ways ;-)
Perhaps it's the configure options you're using which are different (I can
recall more than one bug which you found that I didn't until I tried with
the same options you use).
> got:
>
> #0 0x401d3961 in __kill () at __kill:-1
> #1 0x401d36eb in *__GI_raise (sig=6)
> at ../linuxthreads/sysdeps/unix/sysv/linux/raise.c:33
> #2 0x401d4e87 in *__GI_abort () at ../sysdeps/generic/abort.c:88
> #3 0x0807169f in FatalProblem (sig=11)
> at /usr/local/src/networking/http/browsers/lynx2-8-5/src/LYMain.c:4259
> #4 <signal handler called>
> #5 blank_lines (text=0x8279bd0, newlines=1)
> at /usr/local/src/networking/http/browsers/lynx2-8-5/src/GridText.c:3662
I suspect this loop is what's broken - no null-pointer check for 'line':
while (line != text->last_line &&
HText_TrueEmptyLine(line, text, FALSE)) {
if (newlines == 0)
break;
newlines--; /* Don't bother: already blank */
line = line->prev;
}
Though the real problem may be someplace else (thinking of the textarea
bug, which exposed a couple of missing null-pointer checks).
--
Thomas E. Dickey
http://dickey.his.com
ftp://dickey.his.com
; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden