On Mon, 23 Dec 2024 11:15:26 +0100 Jonas Smedegaard <jo...@jones.dk> wrote: > Quoting Peter Green (2024-12-18 01:38:19) > > Package: sccache > > Version: 0.9.0-1 > > > > I am currently looking at updating rust-backtrace and related packages > > to the latest version, backtrace itself is not semver-breaking but > > a number of it's dependencies are. > > > > backtrace 0.3.69 -> 0.3.74 > > addr2line 0.21.4 -> 0.24.2 > > object 0.32.3 -> 0.36.5 > > gimli 0.28.1 -> 0.31.1 > > ruzstd 0.5.0 -> 0.7.3 > > > > sccache in particular depends on rust-object. > > > > After bumping the dependency I got a failure to find > > macho::FatHeader::parse , checking the upstream changelog I found > > the following entry. > > > > Replaced macho::FatHeader::parse, macho::FatHeader::parse_arch32, > > and macho::FatHeader::parse_arch64 with read::macho::MachOFatFile, > > read::macho::MachOFatFile32 and read::macho::MachOFatFile64. #623 > > > > After doing some exploration of the new API there doesn't seem to > > be a way to try to parse a FatHeader without knowing before hand > > if it is the 32 or 64 bit version. So I made the code try both > > one after the other (and if both fail, it assumes it's not a > > fat binary). > > > > After applying the patch, I got the same number of test failures > > as reported in bug 1090040 > > Thanks for the bugreport, and the (later) patch. > > I am with weak network access the next week, and this seems of lesser > priority due to sccache already broken for other reasons. > > If you prefer faster processing, then either nudge me some more or do a > zero-day NMU, which is fine by me given the narrow changes isolated to > the provided patch.
Jonas, if you don't mind me doing what an Uploader does, I'll help apply the patch, along with fixing the "other reasons" (i.e. #1090040 FTBFS due to a test assuming presence of CARGO_HOME). -- Sdrager, Blair Noctis
OpenPGP_signature.asc
Description: OpenPGP digital signature
