You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@subversion.apache.org by Tom Bing <ca...@gmail.com> on 2012/04/28 00:14:53 UTC
Error in nitial Run of basic_tests.py
I have a Subversion 1.6.15 repository set up and working on an Intel
64-bit RHEL5 Linux platform. I have been trying unsuccessfully to run
the command-line test suite, using
subversion/tests/cmdline/basic_tests.py. The version of Python on
this server is 2.4.3. This script runs normally with the --list
option, and produces a list of 49 test numbers and descriptions on
stdout. When I run the script to perform all tests
(./basic_tests.py), I get nothing on stdout and this on stderr:
Traceback (most recent call last):
File "./basic_tests.py", line 2550, in ?
svntest.main.run_tests(test_
list)
File "/home/tb3798/subversion-1.6.15/subversion/tests/cmdline/svntest/main.py",
line 1607, in run_tests
actions.setup_pristine_repository()
File "/home/tb3798/subversion-1.6.15/subversion/tests/cmdline/svntest/actions.py",
line 44, in setup_pristine_repository
main.create_repos(main.pristine_dir)
File "/home/tb3798/subversion-1.6.15/subversion/tests/cmdline/svntest/main.py",
line 689, in create_repos
path, *opts)
File "/home/tb3798/subversion-1.6.15/subversion/tests/cmdline/svntest/main.py",
line 356, in run_command
None, *varargs)
File "/home/tb3798/subversion-1.6.15/subversion/tests/cmdline/svntest/main.py",
line 500, in run_command_stdin
stdin_lines,
File "/home/tb3798/subversion-1.6.15/subversion/tests/cmdline/svntest/main.py",
line 471, in spawn_process
infile, outfile, errfile, kid = open_pipe([command] + list(varargs))
File "/home/tb3798/subversion-1.6.15/subversion/tests/cmdline/svntest/main.py",
line 418, in open_pipe
close_fds=not windows)
File "/usr/lib64/python2.4/subprocess.py", line 550, in __init__
errread, errwrite)
File "/usr/lib64/python2.4/subprocess.py", line 996, in _execute_child
raise child_exception
OSError: [Errno 2] No such file or directory
I thought this was some kind of configuration error or missing
environment variable, but I don't find anything in the two README
files for testing that covers this.
Thanks for your help,
Tom Bing
Re: Error in nitial Run of basic_tests.py
Posted by Daniel Shahaf <d....@daniel.shahaf.name>.
As to "how could I know this" --- *_tests.py take a --help option,
which lists --bin in its output.
Re: Error in nitial Run of basic_tests.py
Posted by Daniel Shahaf <d....@daniel.shahaf.name>.
*_tests.py look by default for svn in ../../svn/svn --- i.e., build
outputs. Are you running from a built source tree? If not, pass
--bin=/usr/bin.
Re: Error in nitial Run of basic_tests.py
Posted by Daniel Shahaf <da...@elego.de>.
Tom Bing wrote on Tue, May 01, 2012 at 20:04:41 -0400:
> I appreciate your advice, I'll check the SVNKit lists.
>
> Judging from the access timestamps (ls -lu jsvn*) the "jsvnadmin"
> script was the only one that had been accessed when this
> basic_tests.py run failed.
Not surprising; 'svnadmin create' is the first command the harness runs.
Re: Error in nitial Run of basic_tests.py
Posted by Tom Bing <ca...@gmail.com>.
I appreciate your advice, I'll check the SVNKit lists.
Judging from the access timestamps (ls -lu jsvn*) the "jsvnadmin"
script was the only one that had been accessed when this
basic_tests.py run failed. This run, if successful, would have gone
through all the tests.
Regards,
Tom
On Tue, May 1, 2012 at 2:34 PM, Daniel Shahaf <da...@elego.de> wrote:
> Tom Bing wrote on Tue, May 01, 2012 at 12:04:03 -0400:
>> I am running some tests with basic_tests.py to verify the operation of
>> the SVNKit command-line client scripts. I encountered an error (Errno
>> 8) when I ran basic_tests.py with the --use-jsvn
>
> SVNKit questions should be asked on the SVNKit mailing lists.
>
>> and --bin= options, as shown by the terminal session below.
> ..
>> [tom@lorien ~]$ ls $SVNKIT_HOME/bin
>> jsvn jsvndumpfilter jsvnsetup.openvms jsvnversion.bat
>> jsvnadmin jsvndumpfilter.bat jsvnsync
>> jsvnadmin.bat jsvnlook jsvnsync.bat
>> jsvn.bat jsvnlook.bat jsvnversion
>> [tom@lorien ~]$ cd subversion-1.6.15
>> [tom@lorien cmdline]$ ./basic_tests.py --use-jsvn --bin=$SVNKIT_HOME/bin
>> OSError: [Errno 8] Exec format error
>
> I assume that svntest tries gets that error upon attempting to run
> system("$SVNKIT_HOME/bin/jsvn") (aka os.system() in Python).
>
> Beyond that, This does not seem to be a bug in the test harness itself, so please
> follow up on the SVNKit lists. Good luck!
>
> Daniel
Re: Error in nitial Run of basic_tests.py
Posted by Daniel Shahaf <da...@elego.de>.
Tom Bing wrote on Tue, May 01, 2012 at 12:04:03 -0400:
> I am running some tests with basic_tests.py to verify the operation of
> the SVNKit command-line client scripts. I encountered an error (Errno
> 8) when I ran basic_tests.py with the --use-jsvn
SVNKit questions should be asked on the SVNKit mailing lists.
> and --bin= options, as shown by the terminal session below.
..
> [tom@lorien ~]$ ls $SVNKIT_HOME/bin
> jsvn jsvndumpfilter jsvnsetup.openvms jsvnversion.bat
> jsvnadmin jsvndumpfilter.bat jsvnsync
> jsvnadmin.bat jsvnlook jsvnsync.bat
> jsvn.bat jsvnlook.bat jsvnversion
> [tom@lorien ~]$ cd subversion-1.6.15
> [tom@lorien cmdline]$ ./basic_tests.py --use-jsvn --bin=$SVNKIT_HOME/bin
> OSError: [Errno 8] Exec format error
I assume that svntest tries gets that error upon attempting to run
system("$SVNKIT_HOME/bin/jsvn") (aka os.system() in Python).
Beyond that, This does not seem to be a bug in the test harness itself, so please
follow up on the SVNKit lists. Good luck!
Daniel
Re: Error in nitial Run of basic_tests.py
Posted by Tom Bing <ca...@gmail.com>.
I am running some tests with basic_tests.py to verify the operation of
the SVNKit command-line client scripts. I encountered an error (Errno
8) when I ran basic_tests.py with the --use-jsvn and --bin= options,
as shown by the terminal session below.
[tom@lorien ~]$ pwd
/home/tom
[tom@lorien ~]$ echo $SVNKIT_HOME/bin
/home/tom/svnkit-1.3.7/bin
[tom@lorien ~]$ ls $SVNKIT_HOME/bin
jsvn jsvndumpfilter jsvnsetup.openvms jsvnversion.bat
jsvnadmin jsvndumpfilter.bat jsvnsync
jsvnadmin.bat jsvnlook jsvnsync.bat
jsvn.bat jsvnlook.bat jsvnversion
[tom@lorien ~]$ cd subversion-1.6.15
[tom@lorien subversion-1.6.15]$ cd subversion/tests/cmdline
[tom@lorien cmdline]$ rm -rf svn-test-work
[tom@lorien cmdline]$ ./basic_tests.py --use-jsvn --bin=$SVNKIT_HOME/bin
Traceback (most recent call last):
File "./basic_tests.py", line 2550, in ?
svntest.main.run_tests(test_list)
File "/home/tom/subversion-1.6.15/subversion/tests/cmdline/svntest/main.py",
line 1607, in run_tests
actions.setup_pristine_repository()
File "/home/tom/subversion-1.6.15/subversion/tests/cmdline/svntest/actions.py",
line 44, in setup_pristine_repository
main.create_repos(main.pristine_dir)
File "/home/tom/subversion-1.6.15/subversion/tests/cmdline/svntest/main.py",
line 689, in create_repos
path, *opts)
File "/home/tom/subversion-1.6.15/subversion/tests/cmdline/svntest/main.py",
line 356, in run_command
None, *varargs)
File "/home/tom/subversion-1.6.15/subversion/tests/cmdline/svntest/main.py",
line 500, in run_command_stdin
stdin_lines,
File "/home/tom/subversion-1.6.15/subversion/tests/cmdline/svntest/main.py",
line 471, in spawn_process
infile, outfile, errfile, kid = open_pipe([command] + list(varargs))
File "/home/tom/subversion-1.6.15/subversion/tests/cmdline/svntest/main.py",
line 418, in open_pipe
close_fds=not windows)
File "/usr/lib64/python2.4/subprocess.py", line 550, in __init__
errread, errwrite)
File "/usr/lib64/python2.4/subprocess.py", line 996, in _execute_child
raise child_exception
OSError: [Errno 8] Exec format error
[tom@lorien cmdline]$
[tom@lorien cmdline]$ find svn-test-work -type d
svn-test-work
svn-test-work/repositories
svn-test-work/working_copies
svn-test-work/local_tmp
svn-test-work/local_tmp/repos
svn-test-work/local_tmp/config
[tom@lorien cmdline]$ find svn-test-work -type f
svn-test-work/local_tmp/config/config
svn-test-work/local_tmp/config/servers
[tom@lorien cmdline]$
Thanks for your help.
Tom
Re: Error in nitial Run of basic_tests.py
Posted by Daniel Shahaf <da...@elego.de>.
Daniel Shahaf wrote on Mon, Apr 30, 2012 at 17:13:06 +0300:
> Tom Bing wrote on Mon, Apr 30, 2012 at 09:55:33 -0400:
> > Folks, I did try to study the testing READMEs and the info at
> > http://subversion.apache.org/docs/community-guide/building.html before I
> > posted here. I have recently Googled for sites that mention both
> > "basic_tests.py" and "man page", but without useful results.
> >
> > If there's more in-depth documentation on basic_tests.py somewhere (besides
> > the script itself), I'd welcome some guidance here.
> >
>
> The docs we have are:
>
> - subversion/tests/README
> - subversion/tests/cmdline/README
> - --help output
>
> Other docs that might contain or point to useful info are:
> - README
> - perhaps http://subversion.apache.org/HACKING
(I repeated some of those that you mentioned yourself because I wanted
to give the full/exhaustive list of pointers.)
Re: Error in nitial Run of basic_tests.py
Posted by Daniel Shahaf <da...@elego.de>.
Tom Bing wrote on Mon, Apr 30, 2012 at 09:55:33 -0400:
> Hi Folks,
>
> The earlier guidance was a great help. Specifically, this info from Daniel
> on Apr 28:
>
> > *_tests.py look by default for svn in ../../svn/svn --- i.e., build
> > outputs. Are you running from a built source tree? If not, pass
> > --bin=/usr/bin.
>
> By using --bin=, I was able to run the tests successfully against my 1.6.15
> installation.Daniel also informed me about the --help option. I suppose
> that is a universal enough convention now in Subversion for this type of
> script. However, as an old shell-script coder I was more used to "-h" or
> just letting a script get chatty when wrong parameters are supplied.
>
-h works too. (-randomletter gives an error message that, unlike for
'svn' itself, doesn't mention --help.)
> Nico's posts (and Daniel's) on the subtleties of XFAIL are of interest,
> because testing is probably going to be a big part of my job soon. I am
> going to have to scan this thread a few more times before I fully
> understand it.
>
In addition to the scripts (and C programs) in subversion/tests/, you
may want to look at build/run_tests.py too. Feel free to ignore the
WIMP stuffs on first read.
> Folks, I did try to study the testing READMEs and the info at
> http://subversion.apache.org/docs/community-guide/building.html before I
> posted here. I have recently Googled for sites that mention both
> "basic_tests.py" and "man page", but without useful results.
>
> If there's more in-depth documentation on basic_tests.py somewhere (besides
> the script itself), I'd welcome some guidance here.
>
The docs we have are:
- subversion/tests/README
- subversion/tests/cmdline/README
- --help output
Other docs that might contain or point to useful info are:
- README
- perhaps http://subversion.apache.org/HACKING
If you have concrete suggestions (or patches) regarding how to improve
the tests' documentation, we'll be glad to hear them.
Cheers,
Daniel
> Many thanks,
> Tom
>
> On Sun, Apr 29, 2012 at 10:04 AM, Nico Kadel-Garcia <nk...@gmail.com>
> wrote:
> >
> >
> > On Sun, Apr 29, 2012 at 8:54 AM, Daniel Shahaf <d....@daniel.shahaf.name>
> > wrote:
> >>
> >> Nico Kadel-Garcia wrote on Sun, Apr 29, 2012 at 08:48:27 -0400:
> >> > On Sun, Apr 29, 2012 at 1:31 AM, Daniel Shahaf
> >> > <d....@daniel.shahaf.name>wrote:
> >> >
> >> > > Nico Kadel-Garcia wrote on Sat, Apr 28, 2012 at 21:38:59 -0400:
> >> > >
> >> > > > > SKIP and XFAIL are not errors. FAIL are errors. XPASS are not
> >> > > supposed
> >> > > > > to happen in release tarballs, but if they do they are treated
> >> > > > > like
> >> > > FAIL
> >> > > > > is for the purposes of success/failure. The fact that your 'make
> >> > > check'
> >> > > > > exits non-zero despite having neither FAIL nor XPASS is a bug.
> >> > > > >
> >> > > > That's what I get for just posting the last few lines.
> >> > >
> >> > > Recent code prints FAILs after XFAILs and SKIPs, so I can safely say
> >> > > that either you didn't have any FAILs or you posted the output of an
> >> > > oldish version (1.6.x perhaps?).
> >> > >
> >> >
> >> > 1.6.18, I';m afraid. This is on RHEL 5. Let me post the whole part of
> >> > the
> >> > build where it does the checks: Also note, the "make_check" gut turned
> >> > off
> >> > in the RHEL packages long before I arrived on the schene. The Fedora 17
> >> > packages have the checks enabled, and they can pass on RHEL 6: RHEL 5
> >> > and
> >> > RHEL 4 both fail.
> >> >
> >> > Running all tests in utf8_tests.py [55/71]...FAILURE
> >> > Running all tests in changes-test [68/71]...success
> >>
> >> All the Python tests failed, all the C tests passed. tests.log will
> >> have more information.
> >>
> >> (Though: did it really take one hour? it should take under 10 minutes
> >> without the Python tests)
> >
> >
> > I only measured it with the python tests. (I leave out the ruby and java
> > tests on RHEL 4, for which those components do not compile without
> > ridiculous amounts of extra work due to out of date system components.)
> >
> >
> >>
> >> You can run
> >> % make check TESTS="`echo subversion/tests/cmdline/utf8_tests.py`"
> >> or
> >> % cd subversion/tests/cmdline && ./utf8_tests.py
> >> to skip the C tests (and all-but-one of the Python tests).
> >
> >
> > That one shows:
> >
> > subversion-1.6.18]$ make check TESTS="`echo
> > subversion/tests/cmdline/utf8_tests.py`"
> > Running all tests in utf8_tests.py [1/1]...success
> >
> > At least one test was SKIPPED, checking
> > /builddir/build/BUILD/subversion-1.6.18/tests.log
> > SKIP: utf8_tests.py 1: conversion of paths and logs to/from utf8
> > Summary of test results:
> > 1 tests SKIPPED
> >
> > If I look at basic_test.py, I get a lot of errors. I've not had the time
> or
> > expertise with Python to go debugging these.
> >
> > svn:
> >
> 'file:///builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svn-test-work/repositories/basic_tests-39.other/A/B/F'
> > isn't in the same repository as
> >
> 'file:///builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svn-test-work/repositories/basic_tests-39'
> >
> > Traceback (most recent call last):
> > File
> >
> "/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
> > line 1229, in run
> > rc = self.pred.run(**kw)
> > File
> >
> "/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/testcase.py",
> > line 121, in run
> > return self.func(sandbox)
> > File "./basic_tests.py", line 1975, in basic_rm_urls_multi_repos
> > F_url, C_url, F2_url, C2_url)
> > File
> >
> "/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/actions.py",
> > line 202, in run_and_verify_svn
> > expected_exit, *varargs)
> > File
> >
> "/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/actions.py",
> > line 235, in run_and_verify_svn2
> > exit_code, out, err = main.run_svn(want_err, *varargs)
> > File
> >
> "/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
> > line 579, in run_svn
> > return run_command(svn_binary, error_expected, 0,
> > File
> >
> "/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
> > line 356, in run_command
> > None, *varargs)
> > File
> >
> "/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
> > line 515, in run_command_stdin
> > raise Failure
> > Failure
> > XFAIL: basic_tests.py 39: remotely remove directories from two
> repositories
> > EXCEPTION: SVNUnmatchedError
> >
> > Traceback (most recent call last):
> > File
> >
> "/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
> > line 1229, in run
> > rc = self.pred.run(**kw)
> > File
> >
> "/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/testcase.py",
> > line 121, in run
> > return self.func(sandbox)
> > File "./basic_tests.py", line 2255, in info_nonexisting_file
> > raise svntest.main.SVNUnmatchedError
> > SVNUnmatchedError
> > FAIL: basic_tests.py 41: get info on a file not in the repo
> > ERROR: dump failed: svnadmin: Can't open file
> > 'svn-test-work/local_tmp/repos/format': No such file or directory
> > EXCEPTION: SVNRepositoryCopyFailure
> >
> > Traceback (most recent call last):
> > File
> >
> "/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
> > line 1229, in run
> > rc = self.pred.run(**kw)
> > File
> >
> "/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/testcase.py",
> > line 121, in run
> > return self.func(sandbox)
> > File "./basic_tests.py", line 2266, in
> > basic_relative_url_using_current_dir
> > sbox.build()
> > File
> >
> "/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
> > line 1082, in build
> > if actions.make_repo_and_wc(self, create_wc, read_only):
> > File
> >
> "/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/actions.py",
> > line 1393, in make_repo_and_wc
> > guarantee_greek_repository(sbox.repo_dir)
> > File
> >
> "/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/actions.py",
> > line 119, in guarantee_greek_repository
> > if main.copy_repos(main.pristine_dir, path, 1):
> > File
> >
> "/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
> > line 804, in copy_repos
> > raise SVNRepositoryCopyFailure
> > SVNRepositoryCopyFailure
> > FAIL: basic_tests.py 42: basic relative url target using current dir
> > ERROR: dump failed: svnadmin: Can't open file
> > 'svn-test-work/local_tmp/repos/format': No such file or directory
> >
> >
Re: Error in nitial Run of basic_tests.py
Posted by Tom Bing <ca...@gmail.com>.
Hi Folks,
The earlier guidance was a great help. Specifically, this info from Daniel
on Apr 28:
> *_tests.py look by default for svn in ../../svn/svn --- i.e., build
> outputs. Are you running from a built source tree? If not, pass
> --bin=/usr/bin.
By using --bin=, I was able to run the tests successfully against my 1.6.15
installation.Daniel also informed me about the --help option. I suppose
that is a universal enough convention now in Subversion for this type of
script. However, as an old shell-script coder I was more used to "-h" or
just letting a script get chatty when wrong parameters are supplied.
Nico's posts (and Daniel's) on the subtleties of XFAIL are of interest,
because testing is probably going to be a big part of my job soon. I am
going to have to scan this thread a few more times before I fully
understand it.
Folks, I did try to study the testing READMEs and the info at
http://subversion.apache.org/docs/community-guide/building.html before I
posted here. I have recently Googled for sites that mention both
"basic_tests.py" and "man page", but without useful results.
If there's more in-depth documentation on basic_tests.py somewhere (besides
the script itself), I'd welcome some guidance here.
Many thanks,
Tom
On Sun, Apr 29, 2012 at 10:04 AM, Nico Kadel-Garcia <nk...@gmail.com>
wrote:
>
>
> On Sun, Apr 29, 2012 at 8:54 AM, Daniel Shahaf <d....@daniel.shahaf.name>
> wrote:
>>
>> Nico Kadel-Garcia wrote on Sun, Apr 29, 2012 at 08:48:27 -0400:
>> > On Sun, Apr 29, 2012 at 1:31 AM, Daniel Shahaf
>> > <d....@daniel.shahaf.name>wrote:
>> >
>> > > Nico Kadel-Garcia wrote on Sat, Apr 28, 2012 at 21:38:59 -0400:
>> > >
>> > > > > SKIP and XFAIL are not errors. FAIL are errors. XPASS are not
>> > > supposed
>> > > > > to happen in release tarballs, but if they do they are treated
>> > > > > like
>> > > FAIL
>> > > > > is for the purposes of success/failure. The fact that your 'make
>> > > check'
>> > > > > exits non-zero despite having neither FAIL nor XPASS is a bug.
>> > > > >
>> > > > That's what I get for just posting the last few lines.
>> > >
>> > > Recent code prints FAILs after XFAILs and SKIPs, so I can safely say
>> > > that either you didn't have any FAILs or you posted the output of an
>> > > oldish version (1.6.x perhaps?).
>> > >
>> >
>> > 1.6.18, I';m afraid. This is on RHEL 5. Let me post the whole part of
>> > the
>> > build where it does the checks: Also note, the "make_check" gut turned
>> > off
>> > in the RHEL packages long before I arrived on the schene. The Fedora 17
>> > packages have the checks enabled, and they can pass on RHEL 6: RHEL 5
>> > and
>> > RHEL 4 both fail.
>> >
>> > Running all tests in utf8_tests.py [55/71]...FAILURE
>> > Running all tests in changes-test [68/71]...success
>>
>> All the Python tests failed, all the C tests passed. tests.log will
>> have more information.
>>
>> (Though: did it really take one hour? it should take under 10 minutes
>> without the Python tests)
>
>
> I only measured it with the python tests. (I leave out the ruby and java
> tests on RHEL 4, for which those components do not compile without
> ridiculous amounts of extra work due to out of date system components.)
>
>
>>
>> You can run
>> % make check TESTS="`echo subversion/tests/cmdline/utf8_tests.py`"
>> or
>> % cd subversion/tests/cmdline && ./utf8_tests.py
>> to skip the C tests (and all-but-one of the Python tests).
>
>
> That one shows:
>
> subversion-1.6.18]$ make check TESTS="`echo
> subversion/tests/cmdline/utf8_tests.py`"
> Running all tests in utf8_tests.py [1/1]...success
>
> At least one test was SKIPPED, checking
> /builddir/build/BUILD/subversion-1.6.18/tests.log
> SKIP: utf8_tests.py 1: conversion of paths and logs to/from utf8
> Summary of test results:
> 1 tests SKIPPED
>
> If I look at basic_test.py, I get a lot of errors. I've not had the time
or
> expertise with Python to go debugging these.
>
> svn:
>
'file:///builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svn-test-work/repositories/basic_tests-39.other/A/B/F'
> isn't in the same repository as
>
'file:///builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svn-test-work/repositories/basic_tests-39'
>
> Traceback (most recent call last):
> File
>
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
> line 1229, in run
> rc = self.pred.run(**kw)
> File
>
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/testcase.py",
> line 121, in run
> return self.func(sandbox)
> File "./basic_tests.py", line 1975, in basic_rm_urls_multi_repos
> F_url, C_url, F2_url, C2_url)
> File
>
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/actions.py",
> line 202, in run_and_verify_svn
> expected_exit, *varargs)
> File
>
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/actions.py",
> line 235, in run_and_verify_svn2
> exit_code, out, err = main.run_svn(want_err, *varargs)
> File
>
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
> line 579, in run_svn
> return run_command(svn_binary, error_expected, 0,
> File
>
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
> line 356, in run_command
> None, *varargs)
> File
>
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
> line 515, in run_command_stdin
> raise Failure
> Failure
> XFAIL: basic_tests.py 39: remotely remove directories from two
repositories
> EXCEPTION: SVNUnmatchedError
>
> Traceback (most recent call last):
> File
>
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
> line 1229, in run
> rc = self.pred.run(**kw)
> File
>
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/testcase.py",
> line 121, in run
> return self.func(sandbox)
> File "./basic_tests.py", line 2255, in info_nonexisting_file
> raise svntest.main.SVNUnmatchedError
> SVNUnmatchedError
> FAIL: basic_tests.py 41: get info on a file not in the repo
> ERROR: dump failed: svnadmin: Can't open file
> 'svn-test-work/local_tmp/repos/format': No such file or directory
> EXCEPTION: SVNRepositoryCopyFailure
>
> Traceback (most recent call last):
> File
>
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
> line 1229, in run
> rc = self.pred.run(**kw)
> File
>
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/testcase.py",
> line 121, in run
> return self.func(sandbox)
> File "./basic_tests.py", line 2266, in
> basic_relative_url_using_current_dir
> sbox.build()
> File
>
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
> line 1082, in build
> if actions.make_repo_and_wc(self, create_wc, read_only):
> File
>
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/actions.py",
> line 1393, in make_repo_and_wc
> guarantee_greek_repository(sbox.repo_dir)
> File
>
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/actions.py",
> line 119, in guarantee_greek_repository
> if main.copy_repos(main.pristine_dir, path, 1):
> File
>
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
> line 804, in copy_repos
> raise SVNRepositoryCopyFailure
> SVNRepositoryCopyFailure
> FAIL: basic_tests.py 42: basic relative url target using current dir
> ERROR: dump failed: svnadmin: Can't open file
> 'svn-test-work/local_tmp/repos/format': No such file or directory
>
>
Re: Error in nitial Run of basic_tests.py
Posted by Nico Kadel-Garcia <nk...@gmail.com>.
On Sun, Apr 29, 2012 at 8:54 AM, Daniel Shahaf <d....@daniel.shahaf.name>wrote:
> Nico Kadel-Garcia wrote on Sun, Apr 29, 2012 at 08:48:27 -0400:
> > On Sun, Apr 29, 2012 at 1:31 AM, Daniel Shahaf <d.s@daniel.shahaf.name
> >wrote:
> >
> > > Nico Kadel-Garcia wrote on Sat, Apr 28, 2012 at 21:38:59 -0400:
> > >
> > > > > SKIP and XFAIL are not errors. FAIL are errors. XPASS are not
> > > supposed
> > > > > to happen in release tarballs, but if they do they are treated like
> > > FAIL
> > > > > is for the purposes of success/failure. The fact that your 'make
> > > check'
> > > > > exits non-zero despite having neither FAIL nor XPASS is a bug.
> > > > >
> > > > That's what I get for just posting the last few lines.
> > >
> > > Recent code prints FAILs after XFAILs and SKIPs, so I can safely say
> > > that either you didn't have any FAILs or you posted the output of an
> > > oldish version (1.6.x perhaps?).
> > >
> >
> > 1.6.18, I';m afraid. This is on RHEL 5. Let me post the whole part of the
> > build where it does the checks: Also note, the "make_check" gut turned
> off
> > in the RHEL packages long before I arrived on the schene. The Fedora 17
> > packages have the checks enabled, and they can pass on RHEL 6: RHEL 5 and
> > RHEL 4 both fail.
> >
> > Running all tests in utf8_tests.py [55/71]...FAILURE
> > Running all tests in changes-test [68/71]...success
>
> All the Python tests failed, all the C tests passed. tests.log will
> have more information.
>
> (Though: did it really take one hour? it should take under 10 minutes
> without the Python tests)
>
I only measured it with the python tests. (I leave out the ruby and java
tests on RHEL 4, for which those components do not compile without
ridiculous amounts of extra work due to out of date system components.)
> You can run
> % make check TESTS="`echo subversion/tests/cmdline/utf8_tests.py`"
> or
> % cd subversion/tests/cmdline && ./utf8_tests.py
> to skip the C tests (and all-but-one of the Python tests).
>
That one shows:
subversion-1.6.18]$ make check TESTS="`echo
subversion/tests/cmdline/utf8_tests.py`"
Running all tests in utf8_tests.py [1/1]...success
At least one test was SKIPPED, checking
/builddir/build/BUILD/subversion-1.6.18/tests.log
SKIP: utf8_tests.py 1: conversion of paths and logs to/from utf8
Summary of test results:
1 tests SKIPPED
If I look at basic_test.py, I get a lot of errors. I've not had the time or
expertise with Python to go debugging these.
svn:
'file:///builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svn-test-work/repositories/basic_tests-39.other/A/B/F'
isn't in the same repository as
'file:///builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svn-test-work/repositories/basic_tests-39'
Traceback (most recent call last):
File
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
line 1229, in run
rc = self.pred.run(**kw)
File
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/testcase.py",
line 121, in run
return self.func(sandbox)
File "./basic_tests.py", line 1975, in basic_rm_urls_multi_repos
F_url, C_url, F2_url, C2_url)
File
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/actions.py",
line 202, in run_and_verify_svn
expected_exit, *varargs)
File
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/actions.py",
line 235, in run_and_verify_svn2
exit_code, out, err = main.run_svn(want_err, *varargs)
File
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
line 579, in run_svn
return run_command(svn_binary, error_expected, 0,
File
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
line 356, in run_command
None, *varargs)
File
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
line 515, in run_command_stdin
raise Failure
Failure
XFAIL: basic_tests.py 39: remotely remove directories from two repositories
EXCEPTION: SVNUnmatchedError
Traceback (most recent call last):
File
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
line 1229, in run
rc = self.pred.run(**kw)
File
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/testcase.py",
line 121, in run
return self.func(sandbox)
File "./basic_tests.py", line 2255, in info_nonexisting_file
raise svntest.main.SVNUnmatchedError
SVNUnmatchedError
FAIL: basic_tests.py 41: get info on a file not in the repo
ERROR: dump failed: svnadmin: Can't open file
'svn-test-work/local_tmp/repos/format': No such file or directory
EXCEPTION: SVNRepositoryCopyFailure
Traceback (most recent call last):
File
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
line 1229, in run
rc = self.pred.run(**kw)
File
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/testcase.py",
line 121, in run
return self.func(sandbox)
File "./basic_tests.py", line 2266, in
basic_relative_url_using_current_dir
sbox.build()
File
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
line 1082, in build
if actions.make_repo_and_wc(self, create_wc, read_only):
File
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/actions.py",
line 1393, in make_repo_and_wc
guarantee_greek_repository(sbox.repo_dir)
File
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/actions.py",
line 119, in guarantee_greek_repository
if main.copy_repos(main.pristine_dir, path, 1):
File
"/builddir/build/BUILD/subversion-1.6.18/subversion/tests/cmdline/svntest/main.py",
line 804, in copy_repos
raise SVNRepositoryCopyFailure
SVNRepositoryCopyFailure
FAIL: basic_tests.py 42: basic relative url target using current dir
ERROR: dump failed: svnadmin: Can't open file
'svn-test-work/local_tmp/repos/format': No such file or directory
Re: Error in nitial Run of basic_tests.py
Posted by Daniel Shahaf <d....@daniel.shahaf.name>.
Nico Kadel-Garcia wrote on Sun, Apr 29, 2012 at 08:48:27 -0400:
> On Sun, Apr 29, 2012 at 1:31 AM, Daniel Shahaf <d....@daniel.shahaf.name>wrote:
>
> > Nico Kadel-Garcia wrote on Sat, Apr 28, 2012 at 21:38:59 -0400:
> >
> > > > SKIP and XFAIL are not errors. FAIL are errors. XPASS are not
> > supposed
> > > > to happen in release tarballs, but if they do they are treated like
> > FAIL
> > > > is for the purposes of success/failure. The fact that your 'make
> > check'
> > > > exits non-zero despite having neither FAIL nor XPASS is a bug.
> > > >
> > > That's what I get for just posting the last few lines.
> >
> > Recent code prints FAILs after XFAILs and SKIPs, so I can safely say
> > that either you didn't have any FAILs or you posted the output of an
> > oldish version (1.6.x perhaps?).
> >
>
> 1.6.18, I';m afraid. This is on RHEL 5. Let me post the whole part of the
> build where it does the checks: Also note, the "make_check" gut turned off
> in the RHEL packages long before I arrived on the schene. The Fedora 17
> packages have the checks enabled, and they can pass on RHEL 6: RHEL 5 and
> RHEL 4 both fail.
>
> Running all tests in utf8_tests.py [55/71]...FAILURE
> Running all tests in changes-test [68/71]...success
All the Python tests failed, all the C tests passed. tests.log will
have more information.
(Though: did it really take one hour? it should take under 10 minutes
without the Python tests)
You can run
% make check TESTS="`echo subversion/tests/cmdline/utf8_tests.py`"
or
% cd subversion/tests/cmdline && ./utf8_tests.py
to skip the C tests (and all-but-one of the Python tests).
> At least one test was SKIPPED, checking
> /builddir/build/BUILD/subversion-1.6.18/tests.log
> SKIP: cache-test 2: basic memcache svn_cache test
> SKIP: cache-test 3: memcache svn_cache with very long keys
> At least one test XFAILED, checking
> /builddir/build/BUILD/subversion-1.6.18/tests.log
> XFAIL: fs-test 18: merging commit
> XFAIL: locks-test 9: able to reserve a name (lock non-existent path)
> XFAIL: locks-test 10: directory locks (kinda)
> XFAIL: tree-conflict-data-test 3: detect broken tree conflict data
> Summary of test results:
> 2 tests SKIPPED
> 4 tests XFAILED
> make: *** [check] Error 1
> error: Bad exit status from /var/tmp/rpm-tmp.67349 (%check)
Re: Error in nitial Run of basic_tests.py
Posted by Nico Kadel-Garcia <nk...@gmail.com>.
On Sun, Apr 29, 2012 at 1:31 AM, Daniel Shahaf <d....@daniel.shahaf.name>wrote:
> Nico Kadel-Garcia wrote on Sat, Apr 28, 2012 at 21:38:59 -0400:
>
> > > SKIP and XFAIL are not errors. FAIL are errors. XPASS are not
> supposed
> > > to happen in release tarballs, but if they do they are treated like
> FAIL
> > > is for the purposes of success/failure. The fact that your 'make
> check'
> > > exits non-zero despite having neither FAIL nor XPASS is a bug.
> > >
> > That's what I get for just posting the last few lines.
>
> Recent code prints FAILs after XFAILs and SKIPs, so I can safely say
> that either you didn't have any FAILs or you posted the output of an
> oldish version (1.6.x perhaps?).
>
1.6.18, I';m afraid. This is on RHEL 5. Let me post the whole part of the
build where it does the checks: Also note, the "make_check" gut turned off
in the RHEL packages long before I arrived on the schene. The Fedora 17
packages have the checks enabled, and they can pass on RHEL 6: RHEL 5 and
RHEL 4 both fail.
Executing(%check): /bin/sh -e /var/tmp/rpm-tmp.67349
+ umask 022
+ cd /builddir/build/BUILD
+ cd subversion-1.6.18
+ export LANG=C LC_ALL=C
+ LANG=C
+ LC_ALL=C
+ export
LD_LIBRARY_PATH=/var/tmp/subversion-1.6.18-0.2.el5-root-mockbuild/usr/lib64
+
LD_LIBRARY_PATH=/var/tmp/subversion-1.6.18-0.2.el5-root-mockbuild/usr/lib64
+ export MALLOC_PERTURB_=171 MALLOC_CHECK_=3
+ MALLOC_PERTURB_=171
+ MALLOC_CHECK_=3
+ export LIBC_FATAL_STDERR_=1
+ LIBC_FATAL_STDERR_=1
+ /usr/bin/make check check-swig-pl check-swig-py CLEANUP=yes
malloc: using debugging hooks
malloc: using debugging hooks
malloc: using debugging hooks
Running all tests in auth-test [1/71]...success
Running all tests in cache-test [2/71]...success
Running all tests in checksum-test [3/71]...success
Running all tests in client-test [4/71]...success
Running all tests in compat-test [5/71]...success
Running all tests in config-test [6/71]...success
Running all tests in diff-diff3-test [7/71]...success
Running all tests in dirent_uri-test [8/71]...success
Running all tests in error-test [9/71]...success
Running all tests in fs-pack-test [10/71]...success
Running all tests in fs-test [11/71]...success
Running all tests in hashdump-test [12/71]...success
Running all tests in locks-test [13/71]...success
Running all tests in mergeinfo-test [14/71]...success
Running all tests in opt-test [15/71]...success
Running all tests in opt-test [15/71]...success
Running all tests in path-test [16/71]...success
Running all tests in ra-local-test [17/71]...success
Running all tests in random-test [18/71]...success
Running all tests in repos-test [19/71]...success
Running all tests in revision-test [20/71]...success
Running all tests in skel-test [21/71]...success
Running all tests in stream-test [22/71]...success
Running all tests in string-test [23/71]...success
Running all tests in target-test [24/71]...success
Running all tests in time-test [25/71]...success
Running all tests in translate-test [26/71]...success
Running all tests in tree-conflict-data-test [27/71]...success
Running all tests in utf-test [28/71]...success
Running all tests in window-test [29/71]...success
Running all tests in getopt_tests.py [30/71]...FAILURE
Running all tests in basic_tests.py [31/71]...FAILURE
Running all tests in checkout_tests.py [32/71]...FAILURE
Running all tests in commit_tests.py [33/71]...FAILURE
Running all tests in update_tests.py [34/71]...FAILURE
Running all tests in switch_tests.py [35/71]...FAILURE
Running all tests in prop_tests.py [36/71]...FAILURE
Running all tests in schedule_tests.py [37/71]...FAILURE
Running all tests in log_tests.py [38/71]...FAILURE
Running all tests in copy_tests.py [39/71]...FAILURE
Running all tests in diff_tests.py [40/71]...FAILURE
Running all tests in export_tests.py [41/71]...FAILURE
Running all tests in externals_tests.py [42/71]...FAILURE
Running all tests in merge_tests.py [43/71]...FAILURE
Running all tests in merge_authz_tests.py [44/71]...FAILURE
Running all tests in revert_tests.py [45/71]...FAILURE
Running all tests in mergeinfo_tests.py [46/71]...FAILURE
Running all tests in stat_tests.py [47/71]...FAILURE
Running all tests in trans_tests.py [48/71]...FAILURE
Running all tests in autoprop_tests.py [49/71]...FAILURE
Running all tests in blame_tests.py [50/71]...FAILURE
Running all tests in special_tests.py [51/71]...FAILURE
Running all tests in svnadmin_tests.py [52/71]...FAILURE
Running all tests in svnlook_tests.py [53/71]...FAILURE
Running all tests in svnversion_tests.py [54/71]...FAILURE
Running all tests in utf8_tests.py [55/71]...FAILURE
Running all tests in history_tests.py [56/71]...FAILURE
Running all tests in lock_tests.py [57/71]...FAILURE
Running all tests in cat_tests.py [58/71]...FAILURE
Running all tests in import_tests.py [59/71]...FAILURE
Running all tests in svnsync_tests.py [60/71]...FAILURE
Running all tests in authz_tests.py [61/71]...FAILURE
Running all tests in depth_tests.py [62/71]...FAILURE
Running all tests in svndumpfilter_tests.py [63/71]...FAILURE
Running all tests in changelist_tests.py [64/71]...FAILURE
Running all tests in info_tests.py [65/71]...FAILURE
Running all tests in tree_conflict_tests.py [66/71]...FAILURE
Running all tests in resolved_tests.py [67/71]...FAILURE
Running all tests in changes-test [68/71]...success
Running all tests in fs-base-test [69/71]...success
Running all tests in key-test [70/71]...success
Running all tests in strings-reps-test [71/71]...success
At least one test was SKIPPED, checking
/builddir/build/BUILD/subversion-1.6.18/tests.log
SKIP: cache-test 2: basic memcache svn_cache test
SKIP: cache-test 3: memcache svn_cache with very long keys
At least one test XFAILED, checking
/builddir/build/BUILD/subversion-1.6.18/tests.log
XFAIL: fs-test 18: merging commit
XFAIL: locks-test 9: able to reserve a name (lock non-existent path)
XFAIL: locks-test 10: directory locks (kinda)
XFAIL: tree-conflict-data-test 3: detect broken tree conflict data
Summary of test results:
2 tests SKIPPED
4 tests XFAILED
make: *** [check] Error 1
error: Bad exit status from /var/tmp/rpm-tmp.67349 (%check)
Re: Error in nitial Run of basic_tests.py
Posted by Daniel Shahaf <d....@daniel.shahaf.name>.
Nico Kadel-Garcia wrote on Sat, Apr 28, 2012 at 21:38:59 -0400:
> On Sat, Apr 28, 2012 at 5:27 PM, Daniel Shahaf <d....@daniel.shahaf.name>wrote:
>
> > Nico Kadel-Garcia wrote on Sat, Apr 28, 2012 at 16:56:54 -0400:
> > > If you're up for some testing, there are tools I've submitted to
> > Repoforge
> > > for building 1.6.18 and 1.7.4 over at http://www.github.com/nkadel/ All
> > the
> > > patches and .spec files for building the RPM's, compatble for RHEL 4, 5,
> > > and 6. You could rsync the repository to a new environment, build and
> > > install the new RPM's there, and see if the error still occurs.
> > >
> > > I did have to turn off the "make_check" option for normal RPM
> > compilation,
> > > because it takes about an hour to run and it always fails
> >
> > SKIP and XFAIL are not errors. FAIL are errors. XPASS are not supposed
> > to happen in release tarballs, but if they do they are treated like FAIL
> > is for the purposes of success/failure. The fact that your 'make check'
> > exits non-zero despite having neither FAIL nor XPASS is a bug.
> >
> That's what I get for just posting the last few lines.
Recent code prints FAILs after XFAILs and SKIPs, so I can safely say
that either you didn't have any FAILs or you posted the output of an
oldish version (1.6.x perhaps?).
Re: Error in nitial Run of basic_tests.py
Posted by Nico Kadel-Garcia <nk...@gmail.com>.
On Sat, Apr 28, 2012 at 5:27 PM, Daniel Shahaf <d....@daniel.shahaf.name>wrote:
> Nico Kadel-Garcia wrote on Sat, Apr 28, 2012 at 16:56:54 -0400:
> > If you're up for some testing, there are tools I've submitted to
> Repoforge
> > for building 1.6.18 and 1.7.4 over at http://www.github.com/nkadel/ All
> the
> > patches and .spec files for building the RPM's, compatble for RHEL 4, 5,
> > and 6. You could rsync the repository to a new environment, build and
> > install the new RPM's there, and see if the error still occurs.
> >
> > I did have to turn off the "make_check" option for normal RPM
> compilation,
> > because it takes about an hour to run and it always fails
>
> SKIP and XFAIL are not errors. FAIL are errors. XPASS are not supposed
> to happen in release tarballs, but if they do they are treated like FAIL
> is for the purposes of success/failure. The fact that your 'make check'
> exits non-zero despite having neither FAIL nor XPASS is a bug.
>
That's what I get for just posting the last few lines. It fails on quite a
few Python checks. If someone far more comfortable with Pythin than I am
can help out, I'd love to make sure that the checks are failing due to
things like Python version differences, not due to a failure in the
Subversion tools.
Re: Error in nitial Run of basic_tests.py
Posted by Daniel Shahaf <d....@daniel.shahaf.name>.
Nico Kadel-Garcia wrote on Sat, Apr 28, 2012 at 16:56:54 -0400:
> If you're up for some testing, there are tools I've submitted to Repoforge
> for building 1.6.18 and 1.7.4 over at http://www.github.com/nkadel/ All the
> patches and .spec files for building the RPM's, compatble for RHEL 4, 5,
> and 6. You could rsync the repository to a new environment, build and
> install the new RPM's there, and see if the error still occurs.
>
> I did have to turn off the "make_check" option for normal RPM compilation,
> because it takes about an hour to run and it always fails
SKIP and XFAIL are not errors. FAIL are errors. XPASS are not supposed
to happen in release tarballs, but if they do they are treated like FAIL
is for the purposes of success/failure. The fact that your 'make check'
exits non-zero despite having neither FAIL nor XPASS is a bug.
> with some of the
> odder tests for RHEL 4 and RHEL 5. In particular, for RHEL 5 and the
> subversion-1.7.4 SRPM I've built, building inside a "mock" envronment with
> the "make check" enabled: :
>
> Running all tests in key-test [70/71]...success
> Running all tests in strings-reps-test [71/71]...success
> At least one test was SKIPPED, checking
> /builddir/build/BUILD/subversion-1.6.18/tests.log
> SKIP: cache-test 2: basic memcache svn_cache test
> SKIP: cache-test 3: memcache svn_cache with very long keys
> At least one test XFAILED, checking
> /builddir/build/BUILD/subversion-1.6.18/tests.log
> XFAIL: fs-test 18: merging commit
> XFAIL: locks-test 9: able to reserve a name (lock non-existent path)
> XFAIL: locks-test 10: directory locks (kinda)
> XFAIL: tree-conflict-data-test 3: detect broken tree conflict data
> Summary of test results:
> 2 tests SKIPPED
> 4 tests XFAILED
> make: *** [check] Error 1
> error: Bad exit status from /var/tmp/rpm-tmp.67349 (%check)
Re: Error in nitial Run of basic_tests.py
Posted by Nico Kadel-Garcia <nk...@gmail.com>.
If you're up for some testing, there are tools I've submitted to Repoforge
for building 1.6.18 and 1.7.4 over at http://www.github.com/nkadel/ All the
patches and .spec files for building the RPM's, compatble for RHEL 4, 5,
and 6. You could rsync the repository to a new environment, build and
install the new RPM's there, and see if the error still occurs.
I did have to turn off the "make_check" option for normal RPM compilation,
because it takes about an hour to run and it always fails with some of the
odder tests for RHEL 4 and RHEL 5. In particular, for RHEL 5 and the
subversion-1.7.4 SRPM I've built, building inside a "mock" envronment with
the "make check" enabled: :
Running all tests in key-test [70/71]...success
Running all tests in strings-reps-test [71/71]...success
At least one test was SKIPPED, checking
/builddir/build/BUILD/subversion-1.6.18/tests.log
SKIP: cache-test 2: basic memcache svn_cache test
SKIP: cache-test 3: memcache svn_cache with very long keys
At least one test XFAILED, checking
/builddir/build/BUILD/subversion-1.6.18/tests.log
XFAIL: fs-test 18: merging commit
XFAIL: locks-test 9: able to reserve a name (lock non-existent path)
XFAIL: locks-test 10: directory locks (kinda)
XFAIL: tree-conflict-data-test 3: detect broken tree conflict data
Summary of test results:
2 tests SKIPPED
4 tests XFAILED
make: *** [check] Error 1
error: Bad exit status from /var/tmp/rpm-tmp.67349 (%check)