https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101389
--- Comment #4 from ro at CeBiTec dot Uni-Bielefeld.DE <ro at CeBiTec dot Uni-Bielefeld.DE> --- > --- Comment #3 from Gaius Mulley <gaius at gcc dot gnu.org> --- > I've pushed some changes to gcc/m2/Makefile.in on Tue Jan 11 19:21:06 2022 > +0000 > which fix parallel build errors on stage1 binaries (exposed when -j48 was > used). I've rebuilt gm2 using -j48 and -j64 and all regression test pass on > x86_64 debian gnu/linux. Curious to know if it still fails on ASCII.lo on > Solaris 11. I just tried again and got the same error on Solaris 11.4/x86 with -j48, while Linux/x86_64 with -j64 was fine. Afterwards I found the culprit (I think): initially I'd been using make 4.2.92 (don't ask). After I replaced that by 4.3, the build error went away. Btw., could you please have a look at the patch in Comment 4 of PR modula2/101391? It is necessary to fix about 2000+ testsuite failures per multilib on both Solaris and Linux, making m2 testing very hard. Thanks.