hg-1.4 test, solaris: 5 errors
rupert.thurner
eccc8aacd6f9, 5 tests are failing on opencsw, which is a much better
result than ever before, many thanks!
Failed test-bisect: output changed
Failed test-subrepo: output changed
Failed test-diffstat: output changed and returned error code 2
Failed test-patchbomb: output changed and returned error code 1
Failed test-extension: output changed
the test was done on the http://opencsw.org infrastructure, only
solaris 10, sparc. but the tests i did before with the crew repo for
multiple plattforms always showed the same result no matter which
solaris version, or which processor architecture.
rupert.
_______________________________________________
Mercurial-devel mailing list
Mercuri...@selenic.com
http://selenic.com/mailman/listinfo/mercurial-devel
rupert.thurner
> http://mercurial.pastebin.com/f7f400bdacontains a test of hg-stable,
> eccc8aacd6f9, 5 tests are failing on opencsw, which is a much better
> result than ever before, many thanks!
>
> Failed test-bisect: output changed
> Failed test-subrepo: output changed
> Failed test-diffstat: output changed and returned error code 2
> Failed test-patchbomb: output changed and returned error code 1
> Failed test-extension: output changed
>
> solaris 10, sparc. but the tests i did before with the crew repo for
> multiple plattforms always showed the same result no matter which
> solaris version, or which processor architecture.
>
for the skipped svn tests i am not fully sure, what is the
precondition? because the error message says:
# Skipped test-convert-svn-move: missing feature: subversion client
and admin tools
and the system says:
rupert@build10s:~/hg-stable-10s
$ which svn
/opt/csw/bin/svn
$ which svnadmin
/opt/csw/bin/svnadmin
Benoit Boissinot
> http://mercurial.pastebin.com/f7f400bda contains a test of hg-stable,
> eccc8aacd6f9, 5 tests are failing on opencsw, which is a much better
> result than ever before, many thanks!
>
> Failed test-bisect: output changed
> Failed test-subrepo: output changed
> Failed test-diffstat: output changed and returned error code 2
> Failed test-patchbomb: output changed and returned error code 1
> Failed test-extension: output changed
patchbomb and diffstat fixed by:
99a2096c3d8d | 2009-11-08 17:59 +0100 | benoit: fix test-diffstat on solaris
b39f98be96de | 2009-11-08 18:05 +0100 | benoit: fix test-patchbomb on solaris
subrepo is failing here too, I'm not sure about the two others.
thanks!
Benoit
--
:wq
rupert.thurner
wrote:
> On Sun, Nov 08, 2009 at 08:43:31AM -0800, rupert.thurner wrote:
> > eccc8aacd6f9, 5 tests are failing on opencsw, which is a much better
> > result than ever before, many thanks!
>
> > Failed test-bisect: output changed
> > Failed test-subrepo: output changed
> > Failed test-diffstat: output changed and returned error code 2
> > Failed test-patchbomb: output changed and returned error code 1
> > Failed test-extension: output changed
>
> patchbomb and diffstat fixed by:
> 99a2096c3d8d | 2009-11-08 17:59 +0100 | benoit: fix test-diffstat on solaris
> b39f98be96de | 2009-11-08 18:05 +0100 | benoit: fix test-patchbomb on solaris
>
> subrepo is failing here too, I'm not sure about the two others.
>
> thanks!
many thanks!
on solaris, debian, ubuntu, and i guess on every unixishy os, /bin/
bash exists. contrary to /bin/sh, which is the original bourne shell
on solaris. maybe the tests should use /bin/bash in general?
rupert.
Benoit Boissinot
> On 8 Nov., 18:08, Benoit Boissinot <benoit.boissi...@ens-lyon.org>
> wrote:
> > On Sun, Nov 08, 2009 at 08:43:31AM -0800, rupert.thurner wrote:
> > >http://mercurial.pastebin.com/f7f400bdacontains a test of hg-stable,
> > > eccc8aacd6f9, 5 tests are failing on opencsw, which is a much better
> > > result than ever before, many thanks!
> >
> > > Failed test-bisect: output changed
> > > Failed test-subrepo: output changed
> > > Failed test-diffstat: output changed and returned error code 2
> > > Failed test-patchbomb: output changed and returned error code 1
> > > Failed test-extension: output changed
> >
> > patchbomb and diffstat fixed by:
> > 99a2096c3d8d | 2009-11-08 17:59 +0100 | benoit: fix test-diffstat on solaris
> > b39f98be96de | 2009-11-08 18:05 +0100 | benoit: fix test-patchbomb on solaris
> >
> > subrepo is failing here too, I'm not sure about the two others.
> >
> > thanks!
>
> many thanks!
> on solaris, debian, ubuntu, and i guess on every unixishy os, /bin/
> bash exists. contrary to /bin/sh, which is the original bourne shell
> on solaris. maybe the tests should use /bin/bash in general?
It's usually not a problem.
Subrepo should now be fixed in crew, bisect and extension will need to
be investigated.
For bisect, can you give us the return code when trying to execute a
program that does not exist? (hg expects 127)
regards,
Benoit
--
:wq
rupert.thurner
wrote:
> On Sun, Nov 08, 2009 at 09:24:31AM -0800, rupert.thurner wrote:
> > On 8 Nov., 18:08, Benoit Boissinot <benoit.boissi...@ens-lyon.org>
> > wrote:
> > > On Sun, Nov 08, 2009 at 08:43:31AM -0800, rupert.thurner wrote:
> > > > eccc8aacd6f9, 5 tests are failing on opencsw, which is a much better
> > > > result than ever before, many thanks!
>
> > > > Failed test-bisect: output changed
> > > > Failed test-subrepo: output changed
> > > > Failed test-diffstat: output changed and returned error code 2
> > > > Failed test-patchbomb: output changed and returned error code 1
> > > > Failed test-extension: output changed
>
> > > patchbomb and diffstat fixed by:
> > > 99a2096c3d8d | 2009-11-08 17:59 +0100 | benoit: fix test-diffstat on solaris
> > > b39f98be96de | 2009-11-08 18:05 +0100 | benoit: fix test-patchbomb on solaris
>
> > > subrepo is failing here too, I'm not sure about the two others.
>
> > > thanks!
>
> > many thanks!
> > on solaris, debian, ubuntu, and i guess on every unixishy os, /bin/
> > bash exists. contrary to /bin/sh, which is the original bourne shell
> > on solaris. maybe the tests should use /bin/bash in general?
>
> It's usually not a problem.
>
> Subrepo should now be fixed in crew, bisect and extension will need to
> be investigated.
>
> For bisect, can you give us the return code when trying to execute a
> program that does not exist? (hg expects 127)
>
rupert@build10s:~/hg-stable-10s
$ asdf
-bash: asdf: command not found
rupert@build10s:~/hg-stable-10s
$ echo $?
127
rupert@build10s:~/hg-stable-10s
$ sh
$ asdf
asdf: not found
$ echo $?
1
$
rupert.thurner
> On 8 Nov., 18:32, Benoit Boissinot <benoit.boissi...@ens-lyon.org>
> wrote:
>
> > On Sun, Nov 08, 2009 at 09:24:31AM -0800, rupert.thurner wrote:
> > > On 8 Nov., 18:08, Benoit Boissinot <benoit.boissi...@ens-lyon.org>
> > > wrote:
> > > > On Sun, Nov 08, 2009 at 08:43:31AM -0800, rupert.thurner wrote:
you think a change from /bin/sh to /bin/bash in all tests would be
acceptable to avoid such errors for once and forever?
rupert.
Benoit Boissinot
It probably won't fix this particular error anyway (hg is calling
util.system() to invoke the command), so no.
cheers,
Benoit
--
:wq
Mads Kiilerich
> you think a change from /bin/sh to /bin/bash in all tests would be
> acceptable to avoid such errors for once and forever?
If we go down that route then we could just as well also require
"modern" grep and sed and all the other tools mentioned in the issues on
http://mercurial.selenic.com/wiki/WritingTests
/Mads
Yann E. MORIN
All,
On Sunday 08 November 2009 18:58:01 rupert.thurner wrote:
> you think a change from /bin/sh to /bin/bash in all tests would be
> acceptable to avoid such errors for once and forever?
(Un)fortunately, some systems don't have bash in /bin/bash. Some may not
even have bash at all!
Regards,
Yann E. MORIN.
--
.-----------------.--------------------.------------------.--------------------.
| Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ |
| +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. |
`------------------------------^-------^------------------^--------------------'
rupert.thurner
wrote:
> On Sun, Nov 08, 2009 at 09:58:01AM -0800, rupert.thurner wrote:
> > On 8 Nov., 18:47, "rupert.thurner" <rupert.thur...@gmail.com> wrote:
> > > On 8 Nov., 18:32, Benoit Boissinot <benoit.boissi...@ens-lyon.org>
> > > wrote:
>
> > > > For bisect, can you give us the return code when trying to execute a
> > > > program that does not exist? (hg expects 127)
>
> > > this is the shell as well:
>
> > > rupert@build10s:~/hg-stable-10s
> > > $ asdf
> > > -bash: asdf: command not found
>
> > > rupert@build10s:~/hg-stable-10s
> > > $ echo $?
> > > 127
>
> > > rupert@build10s:~/hg-stable-10s
> > > $ sh
> > > $ asdf
> > > asdf: not found
> > > $ echo $?
> > > 1
> > > $
>
> > you think a change from /bin/sh to /bin/bash in all tests would be
> > acceptable to avoid such errors for once and forever?
>
> It probably won't fix this particular error anyway (hg is calling
> util.system() to invoke the command), so no.
>
of course you are right, sorry for confusion. but subprocess contains
"shell=True". and http://docs.python.org/library/subprocess.html says
this would mean /bin/sh as default shell. this then means bourne shell
on solaris with return code 1, and bash on linux with return code 127.
rupert@build10s:~/hg-stable-10s/tests
$ python
Python 2.6.2 (r262:71600, May 27 2009, 13:06:14) [C] on sunos5
Type "help", "copyright", "credits" or "license" for more information.
>>> import os
>>> os.system('asdf')
sh: asdf: not found
256
>>> import subprocess
>>> subprocess.call('asdf',shell=True)
/bin/sh: asdf: not found
1
rupert.thurner
i am confused by this. even the exit code of os.system differs.
$ python
Python 2.6.4 (r264:75706, Oct 29 2009, 15:38:25)
[GCC 4.4.1] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import os
>>> os.system('asdf')
sh: asdf: not found
32512
>>> import subprocess
>>> subprocess.call('asdf',shell=True)
/bin/sh: asdf: not found
127
would http://docs.python.org/library/errno.html#module-errno be
something portable, or not?
Mads Kiilerich
> i am confused by this. even the exit code of os.system differs.
>
> $ python
> Python 2.6.4 (r264:75706, Oct 29 2009, 15:38:25)
> [GCC 4.4.1] on linux2
> Type "help", "copyright", "credits" or "license" for more information.
>>>> import os
>>>> os.system('asdf')
> sh: asdf: not found
> 32512
>>>> import subprocess
>>>> subprocess.call('asdf',shell=True)
> /bin/sh: asdf: not found
> 127
os.system uses the encoding from os.waitpid where the exitvalue has been
shifted 8 bits.
/Mads
Frank Kingswood
> On 8 Nov., 18:47, "rupert.thurner" <rupert.thur...@gmail.com> wrote:
> you think a change from /bin/sh to /bin/bash in all tests would be
> acceptable to avoid such errors for once and forever?
In my opinion it would be better to make the scripts that run /bin/sh
compatible with a strict POSIX shell, removing bashisms. This is usually not
a big deal.
Not everyone running Linux will have bash as /bin/sh. For example, Debian and
derivatives have dash as /bin/sh to improve boot times and memory use.
Frank
Benoit Boissinot
> rupert.thurner wrote:
> > On 8 Nov., 18:47, "rupert.thurner" <rupert.thur...@gmail.com> wrote:
>
> > you think a change from /bin/sh to /bin/bash in all tests would be
> > acceptable to avoid such errors for once and forever?
>
> In my opinion it would be better to make the scripts that run /bin/sh
> compatible with a strict POSIX shell, removing bashisms. This is usually not
> a big deal.
>
> Not everyone running Linux will have bash as /bin/sh. For example, Debian and
> derivatives have dash as /bin/sh to improve boot times and memory use.
It already contains no bashism since we run it on ubuntu very often. But
maybe there is indeed a way to make it use bash on solaris (and /bin/sh
everywhere else) if it helps.
regards,
Benoit
--
:wq
James C. McPherson
> On Mon, Nov 09, 2009 at 01:53:54PM +0000, Frank Kingswood wrote:
>> rupert.thurner wrote:
>>> On 8 Nov., 18:47, "rupert.thurner" <rupert.thur...@gmail.com> wrote:
>>> you think a change from /bin/sh to /bin/bash in all tests would be
>>> acceptable to avoid such errors for once and forever?
>> In my opinion it would be better to make the scripts that run /bin/sh
>> compatible with a strict POSIX shell, removing bashisms. This is usually not
>> a big deal.
>>
>> Not everyone running Linux will have bash as /bin/sh. For example, Debian and
>> derivatives have dash as /bin/sh to improve boot times and memory use.
>
> It already contains no bashism since we run it on ubuntu very often. But
> maybe there is indeed a way to make it use bash on solaris (and /bin/sh
> everywhere else) if it helps.
Does the test output change if you are using bash v3 rather
than bash v4? We recently saw several issues logged when
build snv_126 was released, because the behaviour of
bash v4 differs from bash v3 in several areas.
James C. McPherson
--
Senior Kernel Software Engineer, Solaris
Sun Microsystems
http://blogs.sun.com/jmcp http://www.jmcp.homeunix.com/blog