On Sun, Mar 15, 2026 at 04:42:15PM +0100, Volker Schlecht wrote: > The attached diff builds nim 2.2.8 and passes portcheck. > > - I think we don't need to pull in two different versions of checksums, > right? I dropped one and things still worked after cp'ing "checksums" to > the > first target directory instead of mv'ing it.
IIRC, the two versions had come from discovering the nimble distfile using a different checkout of checksums. That was from 2.2.6 though. 2.2.8 compiles fine for me as well with your diff so might as well drop the second version. > > - I updated the commit hashes to the ones used in 2.2.8 > > - It needed an update-patches, too > > Testing seems to require a pre-existing installation of nim, and there > are quite a lot of test failures (65) that I didn't look into at all - > Maybe it's better than what we have in tree right now, though ... lraab@, > if you want to commit, that's a very wary ok from me ;-) > Cool, I can do that. One last thing, Lydia/Alfred, do either of you have an interest in taking MAINTAINER? It can be shared across both of you as well. > On 3/14/26 1:00 AM, Lydia Sobot wrote: > > > > ? patches/patch-compiler_options_nim > > > > ? patches/patch-koch_nim > > > > ? patches/patch-lib_std_private_threadtypes_nim > > > > Index: Makefile > > > > =================================================================== > > > > > > You need to 'cvs add' these three patches and then regenerate the diff. > > > I suspect they are needed to make the build work. > > They are, I was referring to their existence with "my previous remarks", > > as they are unchanged (and crucially patches/patch-lib_pure_asyncnet_nim > > should be deleted too), but thanks for the cvs add pointer, here they > > are in this new diff, unchanged from Lucas's. > > However, regarding patches/patch-dist_nimble_src_nimblepkg_options_nim, > > I do think that we should perhaps instead just use the built-in git hash > > override facility to patch in the git hash of the release instead of > > removing it entirely.
