[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: LYNX-DEV devel code / fotemods core dumps
From: |
Bill Schiavo |
Subject: |
Re: LYNX-DEV devel code / fotemods core dumps |
Date: |
Fri, 2 May 1997 00:30:57 -0400 (EDT) |
Hello everyone. For just a moment, I thought I was doing something wrong
with the build of lynx but I see that this is not the case. I also
receive core dumps when I exit the program. I first noticed the problem
with the development code at version 2.7.1ac0.12 and also with just about
all of the Fote binaries. Along with the core dump, I get an error
message that says "IOT Trap/abort" Lynx now exiting with signal 11.
I can get a core dump by going to http://home.netscape.com, and after the
document finishes loading, just quit.
I am running Lynx on Linux version 2.0.29.
Bill
On Thu, 1 May 1997, Klaus Weide wrote:
> Two people have reported core dumps occuring with the latest devel code.
> Hynek Med seems to get them on linux since 2.7.1ac-0.12, in a reliable
> manner, Wayne also reported one (I assume it is the same problem). These
> occur only on exit of Lynx, and seem to be dependent on what sites one has
> visited and other browsing history.
>
> I thought this was a problem with the devel code, but now I got one also,
> running a fotemods binary on solaris. A gdb `where' on the core file
> shows that a HTList_delete() called from LYLynxMaps_free() caused the
> exception.
>
> Hynek did a trace of that function and found that the last element
> of LynxMaps was corrupted. A pointer that should have been NULL
> (to signal end of the list) was not NULL. I have tried to reproduce
> that, but so far failed.
>
> i looked at LYMap.c long and couldn't see anything wrong there.
> Apparently some other module is causing memory corruption.
> I have a suspicion that that other module is LYCookie.c (never liked
> them...), this is just a hunch based on showing that that is the next
> thing in memory as shown by `nm'. I'm not familiar with the cookie
> code, but a look showed two things that seem wrong:
>
> - in LYCookieJar_free, this `if (co)' line seems in the wrong place,
> it seems it will always give FALSE
>
> while (cl) {
> next = cl->next;
> if (co) {
> co = cl->object;
> HTList_removeObject(de->cookie_list, co);
>
> - store_cookie() has the following, this `cookie_list' may never get FREEd
> for the `dump_output_immediately' case. (But I may be wrong.)
>
> if (dump_output_immediately) { /* Non-interactive, can't respond */
> if (cookie_list == NULL)
> cookie_list = HTList_new();
>
> I cannot see how these two things could cause memory corruption (instead
> of just unFREEd memory). But there may be other things wrong in that
> file. Could someone more familiar with the cookie code please have
> another look at it?
>
> Klaus
>
> ;
> ; To UNSUBSCRIBE: Send a mail message to address@hidden
> ; with "unsubscribe lynx-dev" (without the
> ; quotation marks) on a line by itself.
> ;
>
;
; To UNSUBSCRIBE: Send a mail message to address@hidden
; with "unsubscribe lynx-dev" (without the
; quotation marks) on a line by itself.
;