Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 698534 - sys-devel/gcc-6.4.0-r1 - checking for sys/stat.h... .../work/gcc-6.4.0/libsanitizer/sanitizer_common/sanitizer_platform_limits_posix.cc:153:23: fatal error: sys/ustat.h: No such file or directory
Summary: sys-devel/gcc-6.4.0-r1 - checking for sys/stat.h... .../work/gcc-6.4.0/libsan...
Status: RESOLVED DUPLICATE of bug 685348
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: Current packages (show other bugs)
Hardware: All Linux
: Normal normal
Assignee: Gentoo Toolchain Maintainers
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2019-10-25 15:42 UTC by Stephan Litterst
Modified: 2019-10-31 12:08 UTC (History)
1 user (show)

See Also:
Package list:
Runtime testing required: ---

Attachments
build.log as bzip2 file (build.log.bz2,337.00 KB, application/x-bzip)
2019-10-25 15:44 UTC, Stephan Litterst 		Details
output of emerge --info (emerge.info,7.78 KB, text/plain)
2019-10-25 15:45 UTC, Stephan Litterst 		Details
output of emerge -pqv (emerge.pqv,275 bytes, text/plain)
2019-10-25 15:45 UTC, Stephan Litterst 		Details
environment (environment.txt,158.23 KB, text/plain)
2019-10-25 15:45 UTC, Stephan Litterst 		Details
gcc-build-logs.tar.bz2 (gcc-build-logs.tar.bz2,533.02 KB, application/x-bzip-compressed-tar)
2019-10-25 15:46 UTC, Stephan Litterst 		Details
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Stephan Litterst 2019-10-25 15:42:47 UTC 
emerge of gcc 6.4.0 failed due to missing ustat.h

I found this in the build.log:
<snip>
checking for sys/stat.h... /var/tmp/portage/sys-devel/gcc-6.4.0-r1/work/gcc-6.4.0/libsanitizer/sanitizer_common/sanitizer_platform_limits_posix.cc:153:23: fatal error: sys/ust
at.h: No such file or directory
 #include <sys/ustat.h>
                       ^
compilation terminated.
make[4]: *** [Makefile:477: sanitizer_platform_limits_posix.lo] Error 1
<snap>

Reproducible: Always

Steps to Reproduce:
1. try to emerge gcc-6.4.0-r1
2.
3.
Comment 1 Stephan Litterst 2019-10-25 15:44:44 UTC 
Created attachment 593926 [details]
build.log as bzip2 file
Comment 2 Stephan Litterst 2019-10-25 15:45:14 UTC 
Created attachment 593928 [details]
output of emerge --info
Comment 3 Stephan Litterst 2019-10-25 15:45:34 UTC 
Created attachment 593930 [details]
output of emerge -pqv
Comment 4 Stephan Litterst 2019-10-25 15:45:50 UTC 
Created attachment 593932 [details]
environment
Comment 5 Stephan Litterst 2019-10-25 15:46:21 UTC 
Created attachment 593934 [details]
gcc-build-logs.tar.bz2
Comment 6 Sergei Trofimovich (RETIRED) gentoo-dev 2019-10-26 12:34:10 UTC 
I think we have fixed it for both 6.4.0-r5 (unstable) and 6.5.0 (stable). Do they not work for you?
Comment 7 Stephan Litterst 2019-10-27 15:14:12 UTC 
Hi Sergej,

the bug does not affect gcc 6.5.0 - it affects gcc 6.4.0-r1.

Kind regards,
Stephan
Comment 8 Nuno Silva 2019-10-27 15:20:28 UTC 
I can reproduce it with 6.4.0-r1 (I ended up installing 6.5.0). It seems to happen with gcc-4.9.4 as well (4.9.4-r1 builds fine).
Comment 9 Sergei Trofimovich (RETIRED) gentoo-dev 2019-10-27 19:33:13 UTC 
(In reply to Stephan Litterst from comment #7)
> Hi Sergej,
> 
> the bug does not affect gcc 6.5.0 - it affects gcc 6.4.0-r1.

Do you need 6.4.0 specifically? It will go away eventually when 6.5.0 is stable on enough platforms.
Comment 10 Stephan Litterst 2019-10-28 05:16:57 UTC 
(In reply to Sergei Trofimovich from comment #9)
> (In reply to Stephan Litterst from comment #7)
> > Hi Sergej,
> > 
> > the bug does not affect gcc 6.5.0 - it affects gcc 6.4.0-r1.
> 
> Do you need 6.4.0 specifically? It will go away eventually when 6.5.0 is
> stable on enough platforms.

I had to run an emerge --changed-deps world and run into bug 698188 and now into the bug mentioned in this ticket.

An "equery d sys-devel/gcc-6.4.0-r1" shows me this:

 * These packages depend on sys-devel/gcc-6.4.0-r1:
app-crypt/libb2-0.98.1-r1 (openmp ? >=sys-devel/gcc-4.2[openmp])
app-office/libreoffice-bin-6.2.5.2 (>=sys-devel/gcc-8.3.0)
app-portage/portage-utils-0.80 (openmp ? >=sys-devel/gcc-4.2[openmp])
dev-db/mariadb-10.2.22-r1 (>=sys-devel/gcc-3.4.6)
dev-java/icedtea-bin-3.13.0 (>=sys-devel/gcc-8.3.0[multilib?])
games-puzzle/world-of-goo-1.41-r2 (>=sys-devel/gcc-3.4)
media-libs/libmypaint-1.3.0-r1 (openmp ? sys-devel/gcc[openmp])
media-libs/mesa-19.1.7 (opencl ? >=sys-devel/gcc-4.6)
net-libs/webkit-gtk-2.24.4 (>=sys-devel/gcc-6.0)
sys-devel/llvm-8.0.1 (>=sys-devel/gcc-3.0)
sys-libs/glibc-2.29-r2 (>=sys-devel/gcc-6)

The output of "eselect gcc list":
 [1] x86_64-pc-linux-gnu-6.4.0
 [2] x86_64-pc-linux-gnu-8.3.0 *

Hope this helps.
Comment 11 Sergei Trofimovich (RETIRED) gentoo-dev 2019-10-28 07:25:53 UTC 
(In reply to Stephan Litterst from comment #10)
> (In reply to Sergei Trofimovich from comment #9)
> > (In reply to Stephan Litterst from comment #7)
> > > Hi Sergej,
> > > 
> > > the bug does not affect gcc 6.5.0 - it affects gcc 6.4.0-r1.
> > 
> > Do you need 6.4.0 specifically? It will go away eventually when 6.5.0 is
> > stable on enough platforms.
> 
> I had to run an emerge --changed-deps world and run into bug 698188 and now
> into the bug mentioned in this ticket.
> 
> An "equery d sys-devel/gcc-6.4.0-r1" shows me this:
> 
>  * These packages depend on sys-devel/gcc-6.4.0-r1:
> app-crypt/libb2-0.98.1-r1 (openmp ? >=sys-devel/gcc-4.2[openmp])
> app-office/libreoffice-bin-6.2.5.2 (>=sys-devel/gcc-8.3.0)
> app-portage/portage-utils-0.80 (openmp ? >=sys-devel/gcc-4.2[openmp])
> dev-db/mariadb-10.2.22-r1 (>=sys-devel/gcc-3.4.6)
> dev-java/icedtea-bin-3.13.0 (>=sys-devel/gcc-8.3.0[multilib?])
> games-puzzle/world-of-goo-1.41-r2 (>=sys-devel/gcc-3.4)
> media-libs/libmypaint-1.3.0-r1 (openmp ? sys-devel/gcc[openmp])
> media-libs/mesa-19.1.7 (opencl ? >=sys-devel/gcc-4.6)
> net-libs/webkit-gtk-2.24.4 (>=sys-devel/gcc-6.0)
> sys-devel/llvm-8.0.1 (>=sys-devel/gcc-3.0)
> sys-libs/glibc-2.29-r2 (>=sys-devel/gcc-6)
> 
> The output of "eselect gcc list":
>  [1] x86_64-pc-linux-gnu-6.4.0
>  [2] x86_64-pc-linux-gnu-8.3.0 *
> 
> Hope this helps.

If the only reason you have gcc-6 is because portage did not clean it up for you automatically you might want to remove unused deps with 'emerge --depclean'.
Comment 12 Stephan Litterst 2019-10-30 20:16:40 UTC 
(In reply to Sergei Trofimovich from comment #11)
> If the only reason you have gcc-6 is because portage did not clean it up for
> you automatically you might want to remove unused deps with 'emerge
> --depclean'.

I always do an emerge --depclean.
Comment 13 Sergei Trofimovich (RETIRED) gentoo-dev 2019-10-30 21:34:50 UTC 
(In reply to Stephan Litterst from comment #12)
> (In reply to Sergei Trofimovich from comment #11)
> > If the only reason you have gcc-6 is because portage did not clean it up for
> > you automatically you might want to remove unused deps with 'emerge
> > --depclean'.
> 
> I always do an emerge --depclean.

If 'emerge --depclean' still does not remove gcc-6 you'll need to find what retains it. Probable candidate is a world entry, for example:
    $ fgrep gcc /var/lib/portage/world
Comment 14 josef.95 2019-10-30 22:31:50 UTC 
Yes, or try
emerge -av --depclean sys-devel/gcc:6.4.0
it should show the reverse dependencies.
Comment 15 Stephan Litterst 2019-10-31 07:09:59 UTC 
I tried both:

fgrep gcc /var/lib/portage/world
sys-devel/gcc
sys-devel/gcc:6.4.0

and:

emerge -av --depclean sys-devel/gcc:6.4.0

Calculating dependencies... done!
>>> Calculating removal order...

>>> These are the packages that would be unmerged:

 sys-devel/gcc
    selected: 6.4.0-r1 
   protected: none 
     omitted: 8.3.0-r1 

All selected packages: =sys-devel/gcc-6.4.0-r1

>>> 'Selected' packages are slated for removal.
>>> 'Protected' and 'omitted' packages will not be removed.

So I removed gcc:6.4.0 and luckily an emerge -uD --changed-deps world does not want to re-emerge it again.

Strangely equery still shows me packages depending on the old gcc:

equery d sys-devel/gcc-6.4.0-r1
 * These packages depend on sys-devel/gcc-6.4.0-r1:
app-crypt/libb2-0.98.1-r1 (openmp ? >=sys-devel/gcc-4.2[openmp])
app-office/libreoffice-bin-6.2.5.2 (>=sys-devel/gcc-8.3.0)
app-portage/portage-utils-0.80 (openmp ? >=sys-devel/gcc-4.2[openmp])
dev-db/mariadb-10.2.22-r1 (>=sys-devel/gcc-3.4.6)
dev-java/icedtea-bin-3.13.0 (>=sys-devel/gcc-8.3.0[multilib?])
games-puzzle/world-of-goo-1.41-r2 (>=sys-devel/gcc-3.4)
media-libs/libmypaint-1.3.0-r1 (openmp ? sys-devel/gcc[openmp])
media-libs/mesa-19.1.7 (opencl ? >=sys-devel/gcc-4.6)
net-libs/webkit-gtk-2.24.4 (>=sys-devel/gcc-6.0)
sys-devel/llvm-8.0.1 (>=sys-devel/gcc-3.0)
sys-libs/glibc-2.29-r2 (>=sys-devel/gcc-6)

Do you know why?
Comment 16 Sergei Trofimovich (RETIRED) gentoo-dev 2019-10-31 07:29:40 UTC 
(In reply to Stephan Litterst from comment #15)
> I tried both:
> 
> fgrep gcc /var/lib/portage/world
> sys-devel/gcc
> sys-devel/gcc:6.4.0



> Strangely equery still shows me packages depending on the old gcc:
> 
> equery d sys-devel/gcc-6.4.0-r1
>  * These packages depend on sys-devel/gcc-6.4.0-r1:
> app-crypt/libb2-0.98.1-r1 (openmp ? >=sys-devel/gcc-4.2[openmp])
> app-office/libreoffice-bin-6.2.5.2 (>=sys-devel/gcc-8.3.0)
> app-portage/portage-utils-0.80 (openmp ? >=sys-devel/gcc-4.2[openmp])
> dev-db/mariadb-10.2.22-r1 (>=sys-devel/gcc-3.4.6)
> dev-java/icedtea-bin-3.13.0 (>=sys-devel/gcc-8.3.0[multilib?])
> games-puzzle/world-of-goo-1.41-r2 (>=sys-devel/gcc-3.4)
> media-libs/libmypaint-1.3.0-r1 (openmp ? sys-devel/gcc[openmp])
> media-libs/mesa-19.1.7 (opencl ? >=sys-devel/gcc-4.6)
> net-libs/webkit-gtk-2.24.4 (>=sys-devel/gcc-6.0)
> sys-devel/llvm-8.0.1 (>=sys-devel/gcc-3.0)
> sys-libs/glibc-2.29-r2 (>=sys-devel/gcc-6)
> 
> Do you know why?

equery does not tell you that. If we look at glibc as an example:

    sys-libs/glibc-2.29-r2 (>=sys-devel/gcc-6)

it says that glibc depends on any >=gcc-6 be it gcc-6,7 or 8. sys-devel/gcc-6.4.0-r1 happens to match >=gcc-6 as well. equery does not do more thant that. There is no redundancy checks against multislot installs.

*** This bug has been marked as a duplicate of bug 685348 ***
Comment 17 Stephan Litterst 2019-10-31 08:34:42 UTC 
(In reply to Sergei Trofimovich from comment #16)
> 
> equery does not tell you that. If we look at glibc as an example:
> 
>     sys-libs/glibc-2.29-r2 (>=sys-devel/gcc-6)
> 
> it says that glibc depends on any >=gcc-6 be it gcc-6,7 or 8.
> sys-devel/gcc-6.4.0-r1 happens to match >=gcc-6 as well. equery does not do
> more thant that. There is no redundancy checks against multislot installs.
> 

Good to know :-)
Thanks a lot for your help!