On Thu, Nov 05, 2026 at 07:09:29PM +0000, nrk=disroot.org via suckless dev list wrote:
> Hi,
>
> I actually had found ~3 of these issues in dwm back at the end of March when
> I
> had some spare tokens to spare. I didn't read the claude report as it was
> too
> large but I'm trying out my testcases now and can confirm that they no
> longer
> crash or trigger ASan/UBSan on master but is faulty when built on commit
> 44dbc68.
>
> The reason why I didn't report these issues back then was because they
> required
> a "malicious client" which is nonsense because X is authenticated and so you
> shouldn't have any malicious client - and if you do you're already
> compromised.
>
> However now that I think about it, there's still the case of a *buggy*
> client
> that does some nonsense due to bugs (e.g trying to reconfigure itself to
> negative, INT_{MIN/MAX} size). Such bugs in client shouldn't crash `dwm` so
> it's good to make things more resilient. But it should be noted that these
> are
> robustness fix, not security fix.
>
> - NRK
>
Nice it fixed the testcases too and I agree.
--
Kind regards,
Hiltjo
Received on Fri Nov 06 2026 - 10:57:35 CET