| View previous topic :: View next topic |
| Author |
Message |
miskolabor n00b

Joined: 23 Aug 2012 Posts: 7
|
Posted: Thu Aug 23, 2012 1:59 pm Post subject: Problem with emerging gcc 4.5.4 |
|
|
Hello all,
Recently I wanted to perform world update and emerging process stopped during gcc compilation with following error:
--------------------------- ACCESS VIOLATION SUMMARY ---------------------------
LOG FILE "/var/log/sandbox/sandbox-19824.log"
VERSION 1.0
FORMAT: F - Function called
FORMAT: S - Access Status
FORMAT: P - Path as passed to function
FORMAT: A - Absolute Path (not canonical)
FORMAT: R - Canonical Path
FORMAT: C - Command Line
F: open_wr
S: deny
P: /usr/include/DONE
A: /usr/include/DONE
R: /usr/include/DONE
C: touch /usr/include/DONE
F: utimensat
S: deny
P: /usr/include/DONE
A: /usr/include/DONE
R: /usr/include/DONE
C: touch /usr/include/DONE
--------------------------------------------------------------------------------
I did revdep-rebuild, cleaned prelink, but the problem still exists.
Here you can find gcc build logs:
https://www.dropbox.com/s/cru9fc39jt0dnx7/gcc-build-logs.tar.bz2
Have anyone faced this kind of problem?
BR, |
|
| Back to top |
|
 |
epsilon72 Guru


Joined: 20 Sep 2007 Posts: 568
|
Posted: Thu Aug 23, 2012 3:26 pm Post subject: |
|
|
I don't know how to truly fix the problem but in the meantime you can upgrade gcc with
| Code: | | # FEATURES="-sandbox" emerge -1 gcc |
|
|
| Back to top |
|
 |
VoidMage Watchman


Joined: 14 Oct 2006 Posts: 6196
|
Posted: Thu Aug 23, 2012 4:02 pm Post subject: |
|
|
| epsilon72 wrote: | I don't know how to truly fix the problem but in the meantime you can upgrade gcc with
| Code: | | # FEATURES="-sandbox" emerge -1 gcc |
|
Please don't pass advices, that are quite likely to eventually break a system.
As for the problem, seems something on the system is broken: just what this '/usr/include/DONE' should be ?
Also, where's the build log ? Those config.logs might be useful later on, but the first step would be looking at the build log. |
|
| Back to top |
|
 |
epsilon72 Guru


Joined: 20 Sep 2007 Posts: 568
|
Posted: Thu Aug 23, 2012 6:17 pm Post subject: |
|
|
| I only posted that because I get sandbox violations emerging firefox all the time (it's a problem with pgo and nvidia in that case) and that is the workaround. I figured it would work here as well. It doesn't break my system...but in this case this user's system might be broken already if gcc is failing to emerge properly. |
|
| Back to top |
|
 |
Hu Administrator

Joined: 06 Mar 2007 Posts: 22088
|
Posted: Thu Aug 23, 2012 10:47 pm Post subject: |
|
|
| It is always a bug if the build fails when the sandbox is turned on. Disabling sandbox might not cause problems if the prohibited operation is safe, but most operations trapped by the sandbox are not safe to unleash on the system. |
|
| Back to top |
|
 |
epsilon72 Guru


Joined: 20 Sep 2007 Posts: 568
|
Posted: Fri Aug 24, 2012 12:43 am Post subject: |
|
|
| Hu wrote: | | It is always a bug if the build fails when the sandbox is turned on. Disabling sandbox might not cause problems if the prohibited operation is safe, but most operations trapped by the sandbox are not safe to unleash on the system. |
Ah, I see. I guess nvidia+firefox+pgo sandbox disabling doing no damage is more an exception to the rule then. |
|
| Back to top |
|
 |
miskolabor n00b

Joined: 23 Aug 2012 Posts: 7
|
|
| Back to top |
|
 |
miskolabor n00b

Joined: 23 Aug 2012 Posts: 7
|
Posted: Mon Aug 27, 2012 9:03 am Post subject: |
|
|
BTW,
as I was getting through this log, found that it miss header files from oracle client - try to fix that and recompile. |
|
| Back to top |
|
 |
miskolabor n00b

Joined: 23 Aug 2012 Posts: 7
|
Posted: Mon Aug 27, 2012 9:55 am Post subject: |
|
|
It was really silly error two slashes instead of one in file path, corrected, gcc comiled and installed |
|
| Back to top |
|
 |
gkonst n00b

Joined: 27 Aug 2012 Posts: 1
|
Posted: Mon Aug 27, 2012 10:54 am Post subject: |
|
|
Hi,
I have the same problem! How did you solve it? |
|
| Back to top |
|
 |
miskolabor n00b

Joined: 23 Aug 2012 Posts: 7
|
Posted: Mon Aug 27, 2012 11:21 am Post subject: |
|
|
My problem was in oracle client header files - In directory /usr/lib64/oracle/10.2.0.3 I've had symlink with double slashes in front of linked dir: client -> //usr/lib64/oracle/10.2.0.3/client/include
I've corrected it to: client -> /usr/lib64/oracle/10.2.0.3/client/include and everything went fine. If you don't use oracle, upload your gcc's build.log file
BR, |
|
| Back to top |
|
 |
sstar n00b

Joined: 05 Sep 2012 Posts: 2
|
Posted: Wed Sep 05, 2012 12:11 pm Post subject: |
|
|
Hi all,
I have the same problem upgrading gcc from 4.5.3-r2 to 4.5.4.
| Code: | | Portage 2.1.11.9 (default/linux/amd64/10.0/desktop, gcc-4.5.3, glibc-2.15-r2, 3.3.8-gentoo x86_64) |
And the error is:
| Code: | make[8]: Leaving directory `/var/tmp/portage/sys-devel/gcc-4.5.4/work/build/x86_64-pc-linux-gnu/32/libstdc++-v3'
make[7]: Leaving directory `/var/tmp/portage/sys-devel/gcc-4.5.4/work/build/x86_64-pc-linux-gnu/32/libstdc++-v3'
make[6]: Leaving directory `/var/tmp/portage/sys-devel/gcc-4.5.4/work/build/x86_64-pc-linux-gnu/32/libstdc++-v3'
make[5]: Leaving directory `/var/tmp/portage/sys-devel/gcc-4.5.4/work/build/x86_64-pc-linux-gnu/libstdc++-v3'
make[4]: Leaving directory `/var/tmp/portage/sys-devel/gcc-4.5.4/work/build/x86_64-pc-linux-gnu/libstdc++-v3'
make[3]: Leaving directory `/var/tmp/portage/sys-devel/gcc-4.5.4/work/build/x86_64-pc-linux-gnu/libstdc++-v3'
make[2]: Leaving directory `/var/tmp/portage/sys-devel/gcc-4.5.4/work/build/x86_64-pc-linux-gnu/libstdc++-v3'
make[1]: Leaving directory `/var/tmp/portage/sys-devel/gcc-4.5.4/work/build'
/var/tmp/portage/sys-devel/gcc-4.5.4/work/build /var/tmp/portage/sys-devel/gcc-4.5.4/work/gcc-4.5.4
>>> Source compiled.
--------------------------- ACCESS VIOLATION SUMMARY ---------------------------
LOG FILE "/var/log/sandbox/sandbox-31539.log"
VERSION 1.0
FORMAT: F - Function called
FORMAT: S - Access Status
FORMAT: P - Path as passed to function
FORMAT: A - Absolute Path (not canonical)
FORMAT: R - Canonical Path
FORMAT: C - Command Line
F: open_wr
S: deny
P: /usr/include/DONE
A: /usr/include/DONE
R: /usr/include/DONE
C: touch /usr/include/DONE
F: utimensat
S: deny
P: /usr/include/DONE
A: /usr/include/DONE
R: /usr/include/DONE
C: touch /usr/include/DONE
--------------------------------------------------------------------------------
*
* Please include /var/tmp/portage/sys-devel/gcc-4.5.4/work/build/gcc-build-logs.tar.bz2 in your bug report
* |
I can't see any errors in /var/tmp/portage/sys-devel/gcc-4.5.4/work/build/gcc-build-logs.tar.bz2. Last lines from it are:
| Code: | #define HAVE_DECL_SBRK 1
#define HAVE_DECL_STRVERSCMP 0
#define NEED_DECLARATION_CANONICALIZE_FILE_NAME 1
#define HAVE_STDLIB_H 1
#define HAVE_UNISTD_H 1
#define HAVE_GETPAGESIZE 1
#define HAVE_MMAP 1
configure: exit 0 |
Any ideas how to fix this? |
|
| Back to top |
|
 |
miskolabor n00b

Joined: 23 Aug 2012 Posts: 7
|
Posted: Wed Sep 05, 2012 12:36 pm Post subject: |
|
|
Hi there,
This info doesn't say much; change dir to /var/tmp/portage/sys-devel/gcc-4.5.4/work/gcc-4.5.4 and look for build.log which contains informations from building gcc.
BR |
|
| Back to top |
|
 |
sstar n00b

Joined: 05 Sep 2012 Posts: 2
|
Posted: Tue Sep 11, 2012 12:29 pm Post subject: |
|
|
| miskolabor wrote: | Hi there,
This info doesn't say much; change dir to /var/tmp/portage/sys-devel/gcc-4.5.4/work/gcc-4.5.4 and look for build.log which contains informations from building gcc.
BR |
I've already posted info from my box to https://bugs.gentoo.org/show_bug.cgi?id=434180 |
|
| Back to top |
|
 |
miskolabor n00b

Joined: 23 Aug 2012 Posts: 7
|
Posted: Tue Sep 11, 2012 12:55 pm Post subject: |
|
|
i think it's not a gcc bug -
It's the same issue I've had. Look at the gcc build log file:
./fixinc.sh: line 388: cd: /var/tmp/portage/sys-devel/gcc-4.5.4/work/build/gcc/include-fixed/root//usr/lib64/oracle/10.2.0.3/client/include: No such file or directory
| Quote: |
My problem was in oracle client header files - In directory /usr/lib64/oracle/10.2.0.3 I've had symlink with double slashes in front of linked dir: client -> //usr/lib64/oracle/10.2.0.3/client/include
I've corrected it to: client -> /usr/lib64/oracle/10.2.0.3/client/include and everything went fine. If you don't use oracle, upload your gcc's build.log file
|
Check if you have the same situation.
BR, |
|
| Back to top |
|
 |
|