search

ClaireTan0216 / memcached

Automatically exported from code.google.com/p/memcached
0 stars 0 forks source link

Tests fail in perl-test harness for 1.4.10 #241

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
What steps will reproduce the problem?
1. make test

What is the expected output? What do you see instead?

What version of the product are you using? On what operating system?
1.4.10, RHEL 5.4, x86_64
perl 5.8.8 x864_64-linux-thread-multi

Please provide any additional information below.

t/00-startup.........ok 16/18Invalid value for binding protocol: http
 -- should be one of auto, binary, or ascii

t/binary.............NOK 267
#   Failed test 'Expected read length'
#   in t/binary.t at line 577.
#          got: '22'
#     expected: '24'
Use of uninitialized value in addition (+) at t/binary.t line 583.
t/binary.............NOK 271
#   Failed test 'Got proper response magic'
#   in t/binary.t at line 581.
#          got: '115'
#     expected: '129'
t/binary.............NOK 274
#   Failed test 'Got proper response magic'
#   in t/binary.t at line 581.
#          got: '0'
#     expected: '129'

Original issue reported on code.google.com by oth...@gmail.com on 22 Dec 2011 at 4:23

GoogleCodeExporter commented 9 years ago 
Do the tests fail with earlier versions? I can't remember offhand if RHEL 5.4 
still had the buggy libevent library or not... What version is on your system?

Original comment by dorma...@rydia.net on 29 Dec 2011 at 10:32

GoogleCodeExporter commented 9 years ago 
closing due to lack of response. please re-open with more information if you're 
still having troubles.

you might try upgrading the point release of centos as well.

Original comment by dorma...@rydia.net on 10 Jan 2012 at 6:02

GoogleCodeExporter commented 9 years ago 
Do tests fail with earlier versions of what? What version of which is on my 
system?

Original comment by oth...@gmail.com on 10 Jan 2012 at 6:46

GoogleCodeExporter commented 9 years ago 
Earlier versions of memcached? like 1.4.9 or something?

what version of libevent do you have?

Original comment by dorma...@rydia.net on 10 Jan 2012 at 7:18

GoogleCodeExporter commented 9 years ago 
libevent-1.4.13

Same problem in 1.4.9, hanging on the last test.

In 1.4.5, a completely different set of test results fail. It gets past those 
above, but chokes on t/whitespace ("Not a git repository"). Do you need more 
info there?

Original comment by oth...@gmail.com on 10 Jan 2012 at 8:35

GoogleCodeExporter commented 9 years ago 
(Note) Same problem in <memcached> 1.4.9.

Original comment by oth...@gmail.com on 10 Jan 2012 at 8:36

GoogleCodeExporter commented 9 years ago 
I'm sorta clueless as to what could be happening. centos is well tested, 
libevent-1.4.13 should be fine, and I've never seen those sort of failures.

1.4.5 shouldn't be failing at that point either. How exactly are you running 
these installs?

Original comment by dorma...@rydia.net on 10 Jan 2012 at 8:48

GoogleCodeExporter commented 9 years ago 
This is just doing an rpmbuild. It's quite possible the tests are disabled on 
many CentOS systems. The RPM spec is from memcached-1.4.5, but I don't think 
this is an issue. No patches are applied. I'm attaching the SPEC File currently 
being used. It's pretty vanilla.

There's probably a perl dependency or two that is missing from the SPEC file 
and that those missing perl modules go undetected. It's perl 5.8.8.

Original comment by oth...@gmail.com on 10 Jan 2012 at 9:50

Attachments:

GoogleCodeExporter commented 9 years ago 
can you try grabbing the tarball and simply doing a ./configure && make && make 
test ? use script to include the full output of the test run if you can.

I'm sure that centos/redhat don't typically error like that...

Original comment by dorma...@rydia.net on 10 Jan 2012 at 10:20

GoogleCodeExporter commented 9 years ago 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
configure && make && make test on 1.4.10 hangs (for at least 1 hour).

t/00-startup.........ok 16/18Invalid value for binding protocol: http
 -- should be one of auto, binary, or ascii
Number of threads must be greater than 0
t/00-startup.........ok
t/64bit..............ok
t/binary-get.........ok
t/binary-sasl........This server is not built with SASL support.
t/binary-sasl........ok
t/binary.............NOK 267
#   Failed test 'Expected read length'
#   in t/binary.t at line 577.
#          got: '22'
#     expected: '24'
Use of uninitialized value in addition (+) at t/binary.t line 583.
t/binary.............NOK 271
#   Failed test 'Got proper response magic'
#   in t/binary.t at line 581.
#          got: '115'
#     expected: '129'
t/binary.............NOK 274
#   Failed test 'Got proper response magic'
#   in t/binary.t at line 581.
#          got: '0'
#     expected: '129'
Timeout.. killing the process

Original comment by oth...@gmail.com on 12 Jan 2012 at 12:34

GoogleCodeExporter commented 9 years ago 
I re-ran the tests with a non-root user and got the same output.

Original comment by oth...@gmail.com on 12 Jan 2012 at 12:36

GoogleCodeExporter commented 9 years ago 
I re-ran the tests with memcached-1.4.9 as both root and non-root and got the 
same output (an earlier test run may have failed because of a previously hung 
test which never was stopped). The output varies slightly:

Invalid value for binding protocol: http
 -- should be one of auto, binary, or ascii
Number of threads must be greater than 0
t/00-startup.........ok
t/64bit..............ok
t/binary-get.........ok
t/binary-sasl........This server is not built with SASL support.
t/binary-sasl........ok
t/binary.............NOK 267
#   Failed test 'Expected read length'
#   in t/binary.t at line 577.
#          got: '23'
#     expected: '24'
Use of uninitialized value in addition (+) at t/binary.t line 583.
t/binary.............NOK 271
#   Failed test 'Got proper response magic'
#   in t/binary.t at line 581.
#          got: '48'
#     expected: '129'
t/binary.............NOK 274
#   Failed test 'Got proper response magic'
#   in t/binary.t at line 581.
#          got: '0'
#     expected: '129'

Original comment by oth...@gmail.com on 12 Jan 2012 at 12:40

GoogleCodeExporter commented 9 years ago 
I can't seem to get your attachment. can you add it to the ticket, or e-mail it 
again?

also can you try 1.4.11-rc1?

I really can't tell from the output why this would be happening, and I don't 
presently have access to a centos box. I know dozens of people who build rpms 
for the platform just fine, so I suspect something is unique with your system.

Unless 1.4.11-rc1 fixes it, then who knows :/

Original comment by dorma...@rydia.net on 12 Jan 2012 at 7:38

GoogleCodeExporter commented 9 years ago 
GOOD NEWS EVERYONE! We accidentally reproduced your bug, though we didn't know 
it at the time.

1.4.12 will come fixed. That is a test failure specific to centos5 on slow 
hardware (single cpus?). One of our buildbot builders got stuck with it and 
dustin was able to fix it.

closing!

Original comment by dorma...@rydia.net on 26 Jan 2012 at 8:16