Additional note: if I revert the commit 377bd37..., everything works. So there's no doubt at this point that I mislabeled commits during the bisect: the segfault is introduced with "Handle IME input".
On Fri, Feb 8, 2019 at 10:40 PM Jordan Timmerman <skor...@gmail.com> wrote: > > Did a git bisect to find the bad commit, and it was in fact the commit > Hiltjo identified: 377bd37e212b1ec4c03a481245603c6560d0be22, "Handle > IME input". > > The output of 'locale' on my system is attached in "locale.out". It's > all en_US.UTF-8. > > I've fiddled with the code very slightly to try to figure out what > change introduced the segfault, but no luck. Might toy around a bit > more if I feel confident I'm not going to break things. > > On Fri, Feb 8, 2019 at 6:47 AM Laslo Hunhold <d...@frign.de> wrote: > > > > On Thu, 7 Feb 2019 22:59:24 -0600 > > Jordan Timmerman <skor...@gmail.com> wrote: > > > > Hey Jordan, > > > > > I'm pretty confident at this point that the segfault occurs during the > > > call to libx11 XmbLookupString. I've attempted downgrading libx11 from > > > 1.6.7 to 1.6.6 and rebuilding, but this makes no difference. For > > > whatever reason, dmenu 4.8 works, and dmenu 4.9 segfaults on keypress. > > > I might look into it again later, but for now I'll downgrade back to > > > dmenu 4.8 and get on with life. :) > > > > how about going to the 4.8 tag in git and then step by step going > > upwards until you find the commit that is to blame? > > > > Sadly I cannot reproduce the segfault on my system, so any more input > > would be greatly appreciated! > > > > With best regards > > > > Laslo > > > > -- > > Laslo Hunhold <d...@frign.de>
