Closed GoogleCodeExporter closed 9 years ago
Which file do I find the output of the result of make check?
Original comment by immaneni...@gmail.com
on 14 Mar 2013 at 4:31
Attached the config.log file to determine the errors.
Original comment by immaneni...@gmail.com
on 14 Mar 2013 at 4:38
Attachments:
./configure `--prefix=/user/aimmanen/`
The above command will reproduce the result
Original comment by immaneni...@gmail.com
on 14 Mar 2013 at 5:07
Some tests require kernel timing mechanism and 2.6.9 is kind old with some bugs
(glibc for instance shows some rt fails even in a RHEL6 kernel). I just checked
r208 on a recent kernel (3.5) and for x86_64 I didn't see any issue on make
check. Could you test it a more recent kernel/toolchain?
Original comment by zatr...@gmail.com
on 14 Mar 2013 at 6:13
I am using a x86 32 bit system and not a 64 bit system so it should be able to
pass the unit test.
Thanks
Original comment by immaneni...@gmail.com
on 14 Mar 2013 at 8:27
rhel4 is very ancient os. Are you really sure you need present day tcmalloc to
work on OS 8 years old ?
Original comment by alkondratenko
on 15 Mar 2013 at 5:57
I just tested r208 on a recent kernel (3.5) for x86 32-bits and all tests
passed. As I said before I believe the failures is due some issues with your
environment (kernel/toolchain) being too ancient.
Original comment by zatr...@gmail.com
on 15 Mar 2013 at 1:56
Thanks, when it was mentioned that it worked with Kernel 3.5 so which RHEL
release did you use?
Original comment by immaneni...@gmail.com
on 15 Mar 2013 at 3:13
Ok, keep in mind I built r208. For the test I used:
Ubuntu 12.10 (kernel 3.5.0, GCC 4.7.2, libunwind 1.0.1)
- 32bits: no issues
- 64bits: no issues
Ubuntu 12.04 (kernel 3.2.0, GCC 4.6.3)
- 32bits: no issues
RHEL6.4 (kernel 2.6.32, GCC 4.4.7)
- 32bits: 1 issue
Testing ./heap-checker_unittest with HEAP_CHECKER_TEST_NO_THREADS=1 ... FAIL
Test was taking unexpectedly long time to run and so we aborted it.
Try the test case manually or raise the timeout from 120
to distinguish test slowness from a real problem.
Output from failed run:
---
WARNING: Perftools heap leak checker is active -- Performance may suffer
Adding pthread-specifics for thread 4152043776 pid 9433
Adding pthread-specifics for thread 4152043776 pid 9433
The testcase seems to entered in a deadlock and since the toolchain is quite
ancient (gcc 4.4, glibc 2.12) I'm almost sure it is some libc issue. I won't
bother look into it.
And I don't think it pays oof the effort to keep investigating issues in old
distros. Since it is working on newer ones with different toolsets, the issue
is likely to be a issue with the old distro itself. If you fell inclined, you
may work on old distro to try debug the issue; but I rather focus on newer ones.
Original comment by zatr...@gmail.com
on 15 Mar 2013 at 3:58
Thank you very much for your reply.
I had tried on a newer version 2.6.18-92.1.18.el5PAE and it passed.
All the unit tests passed showing my configuration of the gperftools were good.
Now I have to update my code to make it compile with gcc version 4.1.2 20080704
before I can proceed using gperftools.
Thanks
Original comment by immaneni...@gmail.com
on 15 Mar 2013 at 4:33
Issue 509 has been merged into this issue.
Original comment by alkondratenko
on 6 Jul 2013 at 10:34
Just tried centos 4.8 i386 inside VM and all tests pass.
Original comment by alkondratenko
on 6 Jul 2013 at 11:34
consider testing 2.1
Original comment by alkondratenko
on 14 Sep 2013 at 8:43
Original issue reported on code.google.com by
immaneni...@gmail.com
on 14 Mar 2013 at 4:30