Bug 710448 – dev-lisp/roswell-19.12.13.103 : fails to build with -fno-common or gcc-10

• Product: Gentoo Linux
• Status: RESOLVED OBSOLETE

[Description Toralf Förster 2020-02-22 03:29:11 UTC]

x86_64-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I..    "-D ROSWELL_HTML_TEST" -O2 -pipe -march=native -fno-common -c -o html_sbcl-proc-opt.o `test -f 'proc-opt.c' || echo './'`proc-opt.c
x86_64-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I..    "-D ROSWELL_HTML_TEST" -O2 -pipe -march=native -fno-common -c -o html_sbcl-manual.o `test -f 'manual.c' || echo './'`manual.c
x86_64-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I..    "-D ROSWELL_HTML_TEST" -O2 -pipe -march=native -fno-common -c -o html_sbcl-manual_windows.o `test -f 'manual_windows.c' || echo './'`manual_windows.c
x86_64-pc-linux-gnu-gcc "-D ROSWELL_HTML_TEST" -O2 -pipe -march=native -fno-common  -Wl,-O1 -Wl,--as-needed -o html_sbcl html_sbcl-html.o html_sbcl-html-sbcl-bin.o html_sbcl-main.o html_sbcl-opt.o html_sbcl-util.o html_sbcl-util_windows.o html_sbcl-util-list.o html_sbcl-util-string.o html_sbcl-util-dir.o html_sbcl-util-dir_windows.o html_sbcl-util-file.o html_sbcl-util-file_windows.o html_sbcl-util-system.o html_sbcl-util-system_windows.o html_sbcl-util-cmdline.o html_sbcl-proc-opt.o html_sbcl-manual.o html_sbcl-manual_windows.o  -lcurl 
/usr/lib/gcc/x86_64-pc-linux-gnu/9.2.0/../../../../x86_64-pc-linux-gnu/bin/ld: html_sbcl-main.o:(.bss+0x40): multiple definition of `argv_orig'; html_sbcl-html-sbcl-bin.o:(.bss+0x0): first defined here
collect2: error: ld returned 1 exit status

  -------------------------------------------------------------------

  This is an unstable amd64 chroot image at a tinderbox (==build bot)
  name: 17.1-test-20200215-161616

  -------------------------------------------------------------------

  Please see the tracker bug for details.

gcc-config -l:
 [1] x86_64-pc-linux-gnu-9.2.0 *

clang:
clang version 9.0.1 
Target: x86_64-pc-linux-gnu
Thread model: posix
InstalledDir: /usr/lib/llvm/9/bin
llvm:
9.0.1
Available Python interpreters, in order of preference:
  [1]   python3.8
  [2]   python3.7
  [3]   python3.6
  [4]   python2.7 (fallback)

Available Rust versions:
  [1]   rust-1.41.0 *
java-config:
The following VMs are available for generation-2:

ghc:
The Glorious Glasgow Haskell Compilation System, version 8.0.2

repository:
==> /var/db/repos/gentoo/metadata/timestamp.chk <==
Sat, 22 Feb 2020 01:37:38 +0000

emerge -qpvO dev-lisp/roswell
[ebuild  N    ] dev-lisp/roswell-19.12.13.103

[Comment 1 Toralf Förster 2020-02-22 03:29:15 UTC]

Created attachment 615014 [details]
emerge-info.txt

[Comment 2 Toralf Förster 2020-02-22 03:29:18 UTC]

Created attachment 615016 [details]
dev-lisp:roswell-19.12.13.103:20200222-021722.log

[Comment 3 Toralf Förster 2020-02-22 03:29:21 UTC]

Created attachment 615018 [details]
emerge-history.txt

[Comment 4 Toralf Förster 2020-02-22 03:29:24 UTC]

Created attachment 615020 [details]
environment

[Comment 5 Toralf Förster 2020-02-22 03:29:27 UTC]

Created attachment 615022 [details]
etc.portage.tbz2

[Comment 6 Toralf Förster 2020-02-22 03:29:29 UTC]

Created attachment 615024 [details]
logs.tbz2

[Comment 7 Toralf Förster 2020-02-22 03:29:32 UTC]

Created attachment 615026 [details]
temp.tbz2

[Comment 8 Danny 2020-02-24 16:06:41 UTC]

I could not reproduct it on my system.

Compiler: gcc version 9.2.0 (Gentoo 9.2.0-r4 p5)
Target: x86_64-pc-linux-gnu

Do you have any extra ld configurations?

[Comment 9 Toralf Förster 2020-02-24 18:46:22 UTC]

(In reply to Danny from comment #8)
> Do you have any extra ld configurations?

except  -fno-common ? 
No

[Comment 10 Danny 2020-02-25 02:00:08 UTC]

Created attachment 615672 [details]
Build log on Danny's machine

[Comment 11 Danny 2020-02-25 02:02:03 UTC]

(In reply to Danny from comment #10)
> Created attachment 615672 [details]
> Build log on Danny's machine

It is strange that I still cannot reproduce it.

Related snippets in make.conf file:

CFLAGS="-O2 -pipe -fno-common"
CXXFLAGS="${CFLAGS}"
CHOST="x86_64-pc-linux-gnu"
FEATURES="nostrip buildpkg parallel-fetch"
CPU_FLAGS_X86="aes avx avx2 f16c fma3 mmx mmxext pclmul popcnt sse sse2 sse3 sse4_1 sse4_2 sse4a ssse3"

I noticed that you have something like "html_sbcl-xxx" in your build log, while I do not have it.

Do you have any clue?

[Comment 12 Sam James 2020-07-25 13:24:49 UTC]

(In reply to Danny from comment #11)
> (In reply to Danny from comment #10)
> > Created attachment 615672 [details]
> > Build log on Danny's machine
> 
> It is strange that I still cannot reproduce it.

Note that ago actually hit this in bug 733152 too, although it wasn't the focus there.

My only guess here is that maybe GCC 10 is stricter? Can you reproduce it with GCC 10 in e.g. a chroot?

[Comment 13 Danny 2020-07-29 03:05:28 UTC]

Created attachment 651344 [details]
gcc -v

[Comment 14 Danny 2020-07-29 03:09:16 UTC]

(In reply to Sam James from comment #12)
> (In reply to Danny from comment #11)
> > (In reply to Danny from comment #10)
> > > Created attachment 615672 [details]
> > > Build log on Danny's machine
> > 
> > It is strange that I still cannot reproduce it.
> 
> Note that ago actually hit this in bug 733152 too, although it wasn't the
> focus there.
> 
> My only guess here is that maybe GCC 10 is stricter? Can you reproduce it
> with GCC 10 in e.g. a chroot?

I am using GCC 10 on my machine. Still cannot hit it :-(

[Comment 15 Sam James 2021-03-20 06:09:55 UTC]

I can build it now at least, let's take that as a win!

[Comment 16 Danny 2021-03-20 14:08:32 UTC]

(In reply to Sam James from comment #15)
> I can build it now at least, let's take that as a win!

I am glad it worked for you! :D