GCC 8.1.1 broken in rawhide due to annobin

[Daniel P. Berrangé]

GCC just updated in Fedora rawhide from 8.0.1 to 8.1.1 and is unable to
create executables due to annobin:


configure:6112: gcc -O2 -g -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2
-Wp,-D_GLIBCXX_ASSERTIONS -fexceptions -fstack-protector-strong -grecord-gcc-switches
-specs=/usr/lib/rpm/redhat/redhat-hardened-cc1
-specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -mtune=generic
-fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection  -Wl,-z,relro 
-Wl,-z,now -specs=/usr/lib/rpm/redhat/redhat-hardened-ld conftest.c  >&5
annobin: conftest.c: Error: plugin built for compiler version (8.1.1) but run with
compiler version (8.0.1)
cc1: error: fail to initialize plugin
/usr/lib/gcc/x86_64-redhat-linux/8/plugin/annobin.so

See

 https://kojipkgs.fedoraproject.org//work/tasks/8382/26748382/build.log
 https://bugzilla.redhat.com/show_bug.cgi?id=1574478

Obviously annobin needs rebuilding, but more generally how can be fix this
long term so that future Gcc updates don't continually break due to annobin
mis-match ?

Regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|

[Neal Gompa]

On Thu, May 3, 2018 at 8:03 AM Daniel P. Berrangé <berrange(a)redhat.com&gt;
wrote:

>  GCC just updated in Fedora rawhide from 8.0.1 to 8.1.1 and is unable
> to
>  create executables due to annobin: 

>  configure:6112: gcc -O2 -g -pipe -Wall -Werror=format-security
-Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -fexceptions
-fstack-protector-strong -grecord-gcc-switches
-specs=/usr/lib/rpm/redhat/redhat-hardened-cc1
-specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -mtune=generic
-fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection
   -Wl,-z,relro  -Wl,-z,now -specs=/usr/lib/rpm/redhat/redhat-hardened-ld
conftest.c  >&5
>  annobin: conftest.c: Error: plugin built for compiler version (8.1.1)
> but 
run with compiler version (8.0.1)
>  cc1: error: fail to initialize plugin
/usr/lib/gcc/x86_64-redhat-linux/8/plugin/annobin.so

>  See 
>    
> https://kojipkgs.fedoraproject.org//work/tasks/8382/26748382/build.log
>     https://bugzilla.redhat.com/show_bug.cgi?id=1574478 
>  Obviously annobin needs rebuilding, but more generally how can be fix
> this
>  long term so that future Gcc updates don't continually break due to 
annobin
>  mis-match ? 

Since we don't have reverse dep auto-rebuilding capabilities, just add the
plugin sources to the gcc package and build it in there. It's pretty clear
by this point that the maintainers can't be bothered to make sure that both
gcc and annobin are updated together so that everything doesn't break.

Sorry for the bluntness, but I'm getting tired of the compiler getting
broken because of this plugin.


--
真実はいつも一つ！/ Always, there's only one truth!

[Daniel P. Berrangé]

On Thu, May 03, 2018 at 12:05:55PM +0000, Neal Gompa wrote:
>  On Thu, May 3, 2018 at 8:03 AM Daniel P. Berrangé
> <berrange(a)redhat.com&gt;
>  wrote:
>  
>  > GCC just updated in Fedora rawhide from 8.0.1 to 8.1.1 and is unable to
>  > create executables due to annobin:
>  
>  
>  > configure:6112: gcc -O2 -g -pipe -Wall -Werror=format-security
>  -Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -fexceptions
>  -fstack-protector-strong -grecord-gcc-switches
>  -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1
>  -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -mtune=generic
>  -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection
>     -Wl,-z,relro  -Wl,-z,now -specs=/usr/lib/rpm/redhat/redhat-hardened-ld
>  conftest.c  >&5
>  > annobin: conftest.c: Error: plugin built for compiler version (8.1.1) but
>  run with compiler version (8.0.1)
>  > cc1: error: fail to initialize plugin
>  /usr/lib/gcc/x86_64-redhat-linux/8/plugin/annobin.so
>  
>  > See
>  
>  >    https://kojipkgs.fedoraproject.org//work/tasks/8382/26748382/build.log
>  >    https://bugzilla.redhat.com/show_bug.cgi?id=1574478
>  
>  > Obviously annobin needs rebuilding, but more generally how can be fix this
>  > long term so that future Gcc updates don't continually break due to
>  annobin
>  > mis-match ?
>  
>  
>  Since we don't have reverse dep auto-rebuilding capabilities, just add the
>  plugin sources to the gcc package and build it in there. It's pretty clear
>  by this point that the maintainers can't be bothered to make sure that both
>  gcc and annobin are updated together so that everything doesn't break.
>  
>  Sorry for the bluntness, but I'm getting tired of the compiler getting
>  broken because of this plugin. 
I think we also need some automated test that validates

 "can create executables with standard RPM build flags"

that is a blocking test before gcc RPMs get pushed into repos since this
is a critical package dep for so much of Fedora.

Regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|

[Adam Williamson]

On Thu, 2018-05-03 at 13:12 +0100, Daniel P. Berrangé wrote:
>  On Thu, May 03, 2018 at 12:05:55PM +0000, Neal Gompa wrote:
>  > On Thu, May 3, 2018 at 8:03 AM Daniel P. Berrangé <berrange(a)redhat.com&gt;
>  > wrote:
>  > 
>  > > GCC just updated in Fedora rawhide from 8.0.1 to 8.1.1 and is unable to
>  > > create executables due to annobin:
>  > 
>  > 
>  > > configure:6112: gcc -O2 -g -pipe -Wall -Werror=format-security
>  > 
>  > -Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -fexceptions
>  > -fstack-protector-strong -grecord-gcc-switches
>  > -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1
>  > -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -mtune=generic
>  > -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection
>  >    -Wl,-z,relro  -Wl,-z,now -specs=/usr/lib/rpm/redhat/redhat-hardened-ld
>  > conftest.c  >&5
>  > > annobin: conftest.c: Error: plugin built for compiler version (8.1.1) but
>  > 
>  > run with compiler version (8.0.1)
>  > > cc1: error: fail to initialize plugin
>  > 
>  > /usr/lib/gcc/x86_64-redhat-linux/8/plugin/annobin.so
>  > 
>  > > See
>  > >    https://kojipkgs.fedoraproject.org//work/tasks/8382/26748382/build.log
>  > >    https://bugzilla.redhat.com/show_bug.cgi?id=1574478
>  > > Obviously annobin needs rebuilding, but more generally how can be fix this
>  > > long term so that future Gcc updates don't continually break due to
>  > 
>  > annobin
>  > > mis-match ?
>  > 
>  > 
>  > Since we don't have reverse dep auto-rebuilding capabilities, just add the
>  > plugin sources to the gcc package and build it in there. It's pretty clear
>  > by this point that the maintainers can't be bothered to make sure that both
>  > gcc and annobin are updated together so that everything doesn't break.
>  > 
>  > Sorry for the bluntness, but I'm getting tired of the compiler getting
>  > broken because of this plugin.
>  
>  I think we also need some automated test that validates
>  
>   "can create executables with standard RPM build flags"
>  
>  that is a blocking test before gcc RPMs get pushed into repos since this
>  is a critical package dep for so much of Fedora. 
That could probably be done in %check in the gcc package itself...
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
http://www.happyassassin.net

[Jonathan Wakely]

On 03/05/18 22:23 -0700, Adam Williamson wrote:
> On Thu, 2018-05-03 at 13:12 +0100, Daniel P. Berrangé wrote:
> > I think we also need some automated test that validates
> >
> >  "can create executables with standard RPM build flags"
> >
> > that is a blocking test before gcc RPMs get pushed into repos since this
> > is a critical package dep for so much of Fedora.
>
> That could probably be done in %check in the gcc package itself... 
Only if annobin is moved into the gcc package, otherwise the %check
would always fail when building a new GCC with a different version
number.

If the test is done against packages in koji that haven't been pushed
to a repo yet then it would allow gcc and annobin to get rebuilt
separately, tested together and pushed together.

[Guido Aulisi]

It seems broken in f28 too.

Ciao
Guido

[Jakub Jelinek]

On Fri, May 04, 2018 at 03:42:11PM +0200, Guido Aulisi wrote:
>  It seems broken in f28 too. 
There is a temporary override so that new annobin can be built, it will be
reverted afterwards.

	Jakub

[Fabio Valentini]

On Fri, May 4, 2018, 15:46 Jakub Jelinek <jakub(a)redhat.com&gt; wrote:

>  On Fri, May 04, 2018 at 03:42:11PM +0200, Guido Aulisi wrote:
>  > It seems broken in f28 too.
>
>  There is a temporary override so that new annobin can be built, it will be
>  reverted afterwards.
>
>  
These shenanigans are exactly what could be avoided by building gcc and its
annobin plugin from the same source package, as Neal suggested.

Fabio


>          Jakub
>  _______________________________________________
>  devel mailing list -- devel(a)lists.fedoraproject.org
>  To unsubscribe send an email to devel-leave(a)lists.fedoraproject.org
>