Re: Patch for run-1.3.0-1 core dump

[Max Polk <maxpolk@gmail.com>]

On 2/18/2014 4:16 AM, Corinna Vinschen wrote:
> On Feb 17 17:25, Jon TURNEY wrote:
>>> ...
>>> I went to all the trouble of investigating this, discovering that
>>> run2_freeargv() is double-freeing the last element in newargv because the NULL
>>> terminator isn't moved when the arguments are shifted down over newargv[0],
>>> and writing a patch, before I noticed that we already had one :-(
>>>
>>> --- origsrc/run-1.3.0/src/run.c 2013-07-24 16:26:39.000000000 +0100
>>> +++ src/run-1.3.0/src/run.c     2014-02-17 17:08:49.125000000 +0000
>>> @@ -254,6 +254,7 @@ realMain(int argc, char* argv[])
>>>         free (newargv[0]);
>>>         for (newargc = 1; newargc < argc; newargc++)
>>>            newargv[newargc-1] = newargv[newargc];
>>> +      newargv[argc-1] = 0;
>>>         newargc = argc - 1;
>>>
>>>         /* update execname */
>> There's still something wrong.  I build run with this patch locally,
>> and it seems to fix the issue at first sight.  However, after the
>> child process of run exits, run throws an exception in free(), and
>> the stack looks broken (on 64 bit).  It seems there is a double free
>> or a free of an entirely unrelated address.
> Scratch that.  I managed to fat-finger a one-line patch.  Sorry.
>
> Corinna

Did my earlier patch get included?  I haven't seen a "run" new version yet.

http://www.cygwin.com/ml/cygwin/2013-12/msg00006.html

My patch was the one that properly quote arguments.  Maybe let's start 
with that before putting new stuff underneath it.  From Chuck: "I'll 
roll a new update fairly soon."

http://www.cygwin.com/ml/cygwin/2013-12/msg00045.html


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple