cast to unbreak llvm22 build; however gnushogi appears to be 12 years
dead upstream and there are other type issues so removal might be a good
option (we also had to fix for -fno-common)
Fix #ifdef maze leading to a build failure on (at least) amd64
Also move additional code upper in the file so that we respect the
pseudo order already present, and so that both hunks are contiguous.
Build error reported by tb@
jujutsu: pull in upstream #9610 to disable mimalloc
I was unsure whether I should disable mimalloc for OpenBSD but then thought
I'd go with upstream's decision despite performance properties of mimalloc
not being great on OpenBSD. While mimalloc has been working fine for me in
the last few weeks and is well tested via python, mimalloc turns out to be
completely broken on illumos, at least the way jujutsu uses it. A day after
release upstream therefore decided not to enable it on platforms they don't
cover in CI. So let just not use this code since it will be disabled again
anyway in a month at the latest.
1.32: Stopped looking for titles in STDs because they went away 2026-05-21
as per usual, upstream doesn't do anything even remotely related to releases,
so pretend one exists and work around that.
OK kn@ (MAINTAINER) sthen@