There is a serious bug in my patch that can cause dvtm to crash which
I have fixed.  After the calls to realloc there are calls to calloc
which were still using the cols field instead of maxcols, so there was
a potential buffer overflow (if you only used the tile and monocle
modes you wouldn't run into this).  Anyway, I've attached the updated
patch.

-Niki

On 9/9/10, Niki Yoshiuchi <aplu...@gmail.com> wrote:
> Thanks! :)
>
> And I realized that my patch is fine.  I couldn't compile it at work, but
> then I remembered that I couldn't compile dvtm using c99 at work because
> our
> Linux box has an outdated version of glibc.  I had to switch to c89 and
> edit
> all the for loops.
>
> On Thu, Sep 9, 2010 at 5:20 PM, Nikhilesh S <s.nikhil...@gmail.com> wrote:
>
>> On Thu, Sep 09, 2010 at 02:02:10PM -0700, Suraj Kurapati wrote:
>> > On 9/8/10, Niki Yoshiuchi <aplu...@gmail.com> wrote:
>> > > even a Gnome user should be able to figure out how to fix it.
>> >
>> > Haha, well said, that quote is an instant classic! :-)
>>
>> I have witnessed a classic in the making. :O
>>
>> --
>> Nikhilesh S
>> http://www.nikhilesh.info
>>
>>
>

Attachment: buffer.diff
Description: Binary data

Reply via email to