x86 build is hanging for 8 hours now: │ └─emerge /usr/lib/python-exec/python3.12/emerge =dev-lisp/sbcl-2.4.9 --oneshot --getbinpkg=n --usepkg-exclude=dev-lisp/sbcl --autounmask-keep-keywords=y --autounmask-use=y --autounmask-continue --autounmask-write │ ├─bash -c ansifilter; exec cat │ │ └─ansifilter │ └─python3.12 /usr/lib/portage/python3.12/pid-ns-init 3043893 │ └─python3.12 /usr/lib/portage/python3.12/pid-ns-init 250 250 250 18 0,1,2 /usr/bin/sandbox [dev-lisp/sbcl-2.4.9] sandbox /usr/lib/portage/python3.12/ebuild.sh compile │ └─sandbox /usr/lib/portage/python3.12/ebuild.sh compile │ └─bash /usr/lib/portage/python3.12/ebuild.sh compile │ └─bash /usr/lib/portage/python3.12/ebuild.sh compile │ └─make.sh ./make.sh sh /var/tmp/portage/dev-lisp/sbcl-2.4.9/work/sbcl-binary/run-sbcl.sh --no-sysinit --no-userinit --disable-debugger │ └─sh -x make-config.sh sh /var/tmp/portage/dev-lisp/sbcl-2.4.9/work/sbcl-binary/run-sbcl.sh --no-sysinit --no-userinit --disable-debugger --check-host-lisp │ └─sbcl --core /var/tmp/portage/dev-lisp/sbcl-2.4.9/work/sbcl-binary/output/sbcl.core --no-sysinit --no-userinit --disable-debugger # strace -p 3043930 strace: Process 3043930 attached read(3, ^Cstrace: Process 3043930 detached <detached ...>
Created attachment 907665 [details] build.log
Created attachment 907666 [details] emerge --info
This *seems* similar to #917495 but is not. In #917495, the C compiler does sothething weird and breaks garbage collection: fatal error encountered in SBCL pid 544 tid 544: GC invariant lost, file "immobile-space.c", line 442 Welcome to LDB, a low-level debugger for the Lisp runtime environment. and the debugger waits for a user input forever. There I made a suggestion: can we add < /dev/null in /usr/lib/portage/python*/ebuild-helpers/emake so that emake breaks instantly, not waiting for a user input which will never come? Probably, this would be useful in other similar cases. But here, it seems from the attached log, sbcl does not run at all: fatal error encountered in SBCL pid 38(tid 0xf7c64700): unexpected return from initial thread in main() It is a different fatal error. And why doesn't the build fail immediately? From this message, it does not seem that it waits for a user input. So, what is it doing for 8 hours? I must say I can reproduce neither #917495 nor this but, and cannot debug them.