Trying to build eclipse-photran:
http://koji.fedoraproject.org/koji/getfile?taskID=791458&name=build.log
+ aot-compile-rpm /usr/lib64/gcj-4.3.1/classmap.db aot-compile-rpm: error: /usr/bin/gcj-dbtool -p: unexpected output
?
Orion Poplawski wrote:
Trying to build eclipse-photran:
http://koji.fedoraproject.org/koji/getfile?taskID=791458&name=build.log
- aot-compile-rpm
/usr/lib64/gcj-4.3.1/classmap.db aot-compile-rpm: error: /usr/bin/gcj-dbtool -p: unexpected output
?
It looks like gcj-dbtool printed its result on stderr instead of stdout, so aot-compile-rpm read nothing from it. Did something change in gcj-dbtool?
Cheers, Gary
Gary Benson wrote:
Orion Poplawski wrote:
Trying to build eclipse-photran:
http://koji.fedoraproject.org/koji/getfile?taskID=791458&name=build.log
- aot-compile-rpm
/usr/lib64/gcj-4.3.1/classmap.db aot-compile-rpm: error: /usr/bin/gcj-dbtool -p: unexpected output
?
It looks like gcj-dbtool printed its result on stderr instead of stdout, so aot-compile-rpm read nothing from it. Did something change in gcj-dbtool?
No. If there's a message on stderr, it's an error message. We need to know what that error message is. I just ptrace'd gcj-dbtool in gcj 4.3.1 and got
[pid 23950] write(1, "/usr/lib64/gcj-4.3.1/classmap.db"..., 33/usr/lib64/gcj-4.3.1/classmap.db ) = 33
Andrew.
Andrew Haley wrote:
Gary Benson wrote:
Orion Poplawski wrote:
Trying to build eclipse-photran:
http://koji.fedoraproject.org/koji/getfile?taskID=791458&name=build.log
- aot-compile-rpm
/usr/lib64/gcj-4.3.1/classmap.db aot-compile-rpm: error: /usr/bin/gcj-dbtool -p: unexpected output
?
It looks like gcj-dbtool printed its result on stderr instead of stdout, so aot-compile-rpm read nothing from it. Did something change in gcj-dbtool?
No. If there's a message on stderr, it's an error message. We need to know what that error message is. I just ptrace'd gcj-dbtool in gcj 4.3.1 and got
[pid 23950] write(1, "/usr/lib64/gcj-4.3.1/classmap.db"..., 33/usr/lib64/gcj-4.3.1/classmap.db ) = 33
The only other thing I can think of then is that Python's os.popen() didn't capture stdout properly. There's nothing in aot-compile-rpm itself that would write "/usr/lib64/gcj-4.3.1/classmap.db" to either stdout or stderr.
Cheers, Gary
Gary Benson wrote:
Andrew Haley wrote:
Gary Benson wrote:
Orion Poplawski wrote:
Trying to build eclipse-photran:
http://koji.fedoraproject.org/koji/getfile?taskID=791458&name=build.log
- aot-compile-rpm
/usr/lib64/gcj-4.3.1/classmap.db aot-compile-rpm: error: /usr/bin/gcj-dbtool -p: unexpected output
?
It looks like gcj-dbtool printed its result on stderr instead of stdout, so aot-compile-rpm read nothing from it. Did something change in gcj-dbtool?
No. If there's a message on stderr, it's an error message. We need to know what that error message is. I just ptrace'd gcj-dbtool in gcj 4.3.1 and got
[pid 23950] write(1, "/usr/lib64/gcj-4.3.1/classmap.db"..., 33/usr/lib64/gcj-4.3.1/classmap.db ) = 33
The only other thing I can think of then is that Python's os.popen() didn't capture stdout properly. There's nothing in aot-compile-rpm itself that would write "/usr/lib64/gcj-4.3.1/classmap.db" to either stdout or stderr.
Yeah. We need to strace aot-compile-rpm at this point to see what it's doing. Orion, is there some way I can logon to this box?
Andrew.
Andrew Haley wrote:
Gary Benson wrote:
The only other thing I can think of then is that Python's os.popen() didn't capture stdout properly. There's nothing in aot-compile-rpm itself that would write "/usr/lib64/gcj-4.3.1/classmap.db" to either stdout or stderr.
Yeah. We need to strace aot-compile-rpm at this point to see what it's doing. Orion, is there some way I can logon to this box?
This is probably the popen() broken on xen kernels bug. I'll resubmit my build.
Orion Poplawski wrote:
Andrew Haley wrote:
Gary Benson wrote:
The only other thing I can think of then is that Python's os.popen() didn't capture stdout properly. There's nothing in aot-compile-rpm itself that would write "/usr/lib64/gcj-4.3.1/classmap.db" to either stdout or stderr.
Yeah. We need to strace aot-compile-rpm at this point to see what it's doing. Orion, is there some way I can logon to this box?
This is probably the popen() broken on xen kernels bug.
Ah. This seems very likely! :-)
Andrew.
Andrew Haley wrote:
Orion Poplawski wrote:
Andrew Haley wrote:
Gary Benson wrote:
The only other thing I can think of then is that Python's os.popen() didn't capture stdout properly. There's nothing in aot-compile-rpm itself that would write "/usr/lib64/gcj-4.3.1/classmap.db" to either stdout or stderr.
Yeah. We need to strace aot-compile-rpm at this point to see what it's doing. Orion, is there some way I can logon to this box?
This is probably the popen() broken on xen kernels bug.
Ah. This seems very likely! :-)
Quite :)
Orion Poplawski wrote:
Andrew Haley wrote:
Gary Benson wrote:
The only other thing I can think of then is that Python's os.popen() didn't capture stdout properly. There's nothing in aot-compile-rpm itself that would write "/usr/lib64/gcj-4.3.1/classmap.db" to either stdout or stderr.
Yeah. We need to strace aot-compile-rpm at this point to see what it's doing. Orion, is there some way I can logon to this box?
This is probably the popen() broken on xen kernels bug. I'll resubmit my build.
Build worked today.
Orion Poplawski wrote:
Orion Poplawski wrote:
Andrew Haley wrote:
Gary Benson wrote:
The only other thing I can think of then is that Python's os.popen() didn't capture stdout properly. There's nothing in aot-compile-rpm itself that would write "/usr/lib64/gcj-4.3.1/classmap.db" to either stdout or stderr.
Yeah. We need to strace aot-compile-rpm at this point to see what it's doing. Orion, is there some way I can logon to this box?
This is probably the popen() broken on xen kernels bug. I'll resubmit my build.
Build worked today.
Excellent, thanks.
Cheers, Gary
java-devel@lists.fedoraproject.org
-
Andrew Haley -
Gary Benson -
Orion Poplawski