Another year another version. I understand it may take awhile for it to be supported, but putting this in. Currently the bootstrap fails in stage1 right after unpacking the portage snapshot: 2024-09-21 05:50:37 (8.61 MB/s) - 'portage-20240729.tar.bz2' saved [58898622/58898622] * Unpacking, this may take a while tage1: unpacking Portage tree^G* Setting up some guessed defaults * Your profile is set to /opt/gentoo/var/db/repos/gentoo/profiles/prefix/darwin/macos/15.0/arm64/gcc. ./bootstrap-prefix.sh: line 473: /opt/gentoo/etc/portage/make.profile/package.use: No such file or directory ./bootstrap-prefix.sh: line 482: /opt/gentoo/etc/portage/make.profile/package.unmask: No such file or directory Reproducible: Always
yes, you need to use LATEST_TREE, but then you'll run into the most obscure problems while building - as no longer accepts input from stdin without '-' as argument (I tried working around that locally to get further) - mv aborts (SIGABRT) while moving files
ok, there's some progress here: - profiles are defined - gcc-14.2.0 now compiles - the compiler can produce working programs :) I just pushed the gcc fix, so we can use a snapshot tomorrow morning earliest
That got a lot farther for me, though gcc still failed to build in stage 2. Unless I didn't wait long enough to try. I'll attach the gcc-build-logs.
Created attachment 903762 [details] gcc-build-logs.tar.xz
this is really (not) interesting - my arm64 fails on "Unsupported value of __FLT_EVAL_METHOD__." - my x64 fails on a linker error (undefined symbol of something unwind) - yours fails on "clang: no input files", which I should've fixed in tree
Hmm maybe that was user error. I used LATEST_TREE=1 instead of LATEST_TREE_YES=1. I'll attach the new build logs with the latter. If it matters, this is a first-generation M1 MacBook Pro (2021).
Created attachment 904076 [details] LATEST_TREE_YES=1
A little yay, I managed to finish an arm64-darwin24 bootstrap. With a bit of luck, tomorrow's snapshot will do the trick.
It worked!!! Thanks Fabian! That is the fastest support for a new macOS that I can remember, by far.
Thanks for confirming, it seems to run fine in the tests on older releases too, so closing this bug!
