Created attachment 696993 [details] build.log Both 1.16.2 and 1.16.3 versions of dev-lang/go refuses to build on my x86 system; earlier versions had no issues. The main error I'm gettin is: go tool dist: FAILED: /var/tmp/portage/dev-lang/go-1.16.3/work/go/pkg/tool/linux_386/asm -I /var/tmp/portage/dev-lang/go-1.16.3/temp/go-tool-dist-351048126/internal/cpu -I /var/tmp/portage/dev-lang/go-1.16.3/work/go/pkg/include -D GOOS_linux -D GOARCH_386 -D GOOS_GOARCH_linux_386 -p internal/cpu -gensymabis -o /var/tmp/portage/dev-lang/go-1.16.3/temp/go-tool-dist-351048126/internal/cpu/symabis /var/tmp/portage/dev-lang/go-1.16.3/work/go/src/internal/cpu/cpu.s /var/tmp/portage/dev-lang/go-1.16.3/work/go/src/internal/cpu/cpu_x86.s: signal: illegal instruction (complete build.log attached) The only reference to similar bug in go-lang I was able to find is this: https://groups.google.com/g/golang-nuts/c/SmY82rp7QQU Does the error I'm getting mean that Go developers have broken the support for older non-Intel 32-bit processors in go-1.16.x? My system is using an AMD Athlon XP-processor.
Created attachment 696996 [details] emerge --info
Created attachment 696999 [details] /proc/cpuinfo
My guess would be that it's trying to use SSE2 (which you don't have), not that I have such machine to confirm.
If you have a non-SSE2 386, the following affects you: https://golang.org/doc/go1.16#386 https://golang.org/doc/go1.15#386 Go 1.15 is the last release to support x87-only floating-point hardware (GO386=387). Future releases will require at least SSE2 support on 386, raising Go's minimum GOARCH=386 requirement to the Intel Pentium 4 (released in 2000) or AMD Opteron/Athlon 64 (released in 2003). I am not sure what I can do since this is an upstream decision, so please let me know.
This may be the issue.
(In reply to William Hubbs from comment #4) > I am not sure what I can do since this is an upstream decision, so > please let me know. Well, because it is Google's stupid decision, I don't think you can, or need to do anything. As for me, I need to either stick to go-1.15 for the rest of that old machine's life, or stop using all software needing go.
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=dc46df8c41da1cd54074683f0ca389e92e9a9707 commit dc46df8c41da1cd54074683f0ca389e92e9a9707 Author: William Hubbs <williamh@gentoo.org> AuthorDate: 2021-06-06 19:56:50 +0000 Commit: William Hubbs <williamh@gentoo.org> CommitDate: 2021-06-06 19:56:50 +0000 dev-lang/go: 1.16.5 bump Bug: https://bugs.gentoo.org/779859 Package-Manager: Portage-3.0.18, Repoman-3.0.2 Signed-off-by: William Hubbs <williamh@gentoo.org> dev-lang/go/Manifest | 1 + dev-lang/go/go-1.16.5.ebuild | 189 +++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 190 insertions(+)
Please test go 1.16.5; you should be able to build it. Thanks, William
I am closing this as needinfo because there has been no response. Please report the status of building go 1.16.x and feel free to re-open this bug if it still can't be built. Thanks, William
The machine I was trying to build go-1.16 finally broke down two weeks ago after almost 20 years of service... This heatwave was probably too much for it.
Hello, I'm hit by this bug on my (old) HP server (dual pentium III Coppermine). Neither go 1.16.5 nor 1.16.7 build on this machine. I've attached the build.log to the bug report.
Created attachment 733105 [details] build.log
(In reply to Spekadyon from comment #11) > Hello, > > I'm hit by this bug on my (old) HP server (dual pentium III Coppermine). > Neither go 1.16.5 nor 1.16.7 build on this machine. I've attached the > build.log to the bug report. I think this might be a different bug, actually. Could you file a new one please?
(In reply to Sam James from comment #13) > (In reply to Spekadyon from comment #11) > > Hello, > > > > I'm hit by this bug on my (old) HP server (dual pentium III Coppermine). > > Neither go 1.16.5 nor 1.16.7 build on this machine. I've attached the > > build.log to the bug report. > > I think this might be a different bug, actually. Could you file a new one > please? Yes, I read the bug report too fast, sorry. I'm opening a new one.