You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Jens Peters <jp...@gmx.de> on 2008/06/01 11:36:47 UTC
[PATCH] Let depth-tests-23 pass on Windows for dav|svn and bdb
Hello there,
As mentioned here earlier I see depth-tests failures on my system with the combination svn | dav and bdb. See
http://svn.haxx.se/dev/archive-2008-05/0166.shtml for details. I finally found some time to dig into it and as a result the attached patch fixes the
issue for me.
I know that this is more an ugly workaround than a real fix. But is seems to work fine and at least all tests are green now (which feels actually
pretty good, +1 to release from me :) ). Any chance that this patch or anything similar can be committed? Note thas this patch is made against
1.5-rc8. I would love to make patch against trunk, but first I would like to know if it will be accepted.
Regards, Jens
[[[
Fix a depth-tests-23 failure (WindowsError: Access is denied) on Windows for the combination [svn | dav] and [bdb].
* subversion/tests/cmdline/depth_tests.py
Insert one second sleep to give Windows the time to release the Berkeley DB file handles before recreating the repository.
]]]
RE: [PATCH] Let depth-tests-23 pass on Windows for dav|svn and bdb
Posted by Jens Peters <jp...@gmx.de>.
>If this is specific for Windows, a check for os type (os.name) would be
good,
>along with a comment explaining why we have this as dglasser has pointed
out.
Thanks a lot for your and dglassers comments, very cool! I'll improve the
patch in the weekend, I'm currently busy with real-life work and travelling.
Regards,
Jens
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
RE: [PATCH] Let depth-tests-23 pass on Windows for dav|svn and bdb
Posted by Jens Peters <jp...@gmx.de>.
>If this is specific for Windows, a check for os type (os.name) would be
good,
>along with a comment explaining why we have this as dglasser has pointed
out.
mmh, reading that post I don't think so:
http://svn.haxx.se/dev/archive-2008-06/0151.shtml, though I have this
failure always (with svn or dav).
Regards,
Jens
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: [PATCH] Let depth-tests-23 pass on Windows for dav|svn and bdb
Posted by Jens Peters <jp...@gmx.de>.
Daniel Shahaf wrote:
> Jens, ideally we'd have more than just a "guess"; can you check other
> platforms?
Actually no, sorry, I just have a WinXP and a Vista box with a svn build
environment around here, whereas the XP box is an older laptop where I
got that failure. The Vista machine is much faster and there the test
suite completes successfully without the patch.
The 'guess' comes from Kamesh's report where he has seen the same
failure on a non windows box.
I'm sorry, I cant give you any 'hard' facts, just the fact that the one
second sleep let the test pass on the older machine :(, though I
understand your request of course.
Mark, as far as I remember you had the same strange objections. Do you
have some ideas?
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: [PATCH] Let depth-tests-23 pass on Windows for dav|svn and bdb
Posted by Daniel Shahaf <d....@daniel.shahaf.co.il>.
Ping. (Context: David Glasser said earlier he wasn't sure this is the
best fix.)
See below...
Jens Peters wrote on Sun, 8 Jun 2008 at 14:55 +0200:
> Senthil Kumaran S wrote:
> > If this is specific for Windows, a check for os type (os.name) would be
> > good, along with a comment explaining why we have this as dglasser has
> > pointed out.
> >
>
> I did a rerun on a much faster system and there I didn't got that
> failure, thus I assume this is just a result of my 6 year old laptop ;).
> According to Kamesh's report (
> http://svn.haxx.se/dev/archive-2008-06/0151.shtml) I guess also that
> this is not a windows only problem, but just occurs more often on
> windows. I have attached a new patch with some explanation.
>
Jens, ideally we'd have more than just a "guess"; can you check other
platforms?
>
> [[[
> Fix a depth-tests-23 failure (Access is denied) for the combination
> [bdb] x [svn | dav].
>
> * subversion/tests/cmdline/depth_tests.py
> Insert one second sleep to give slower systems the time to release
> the Berkeley DB shared memory files before recreating the repository.
>
> Patch by: jpeters7677@gmx.de
> ]]]
>
>
>
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: [PATCH] Let depth-tests-23 pass on Windows for dav|svn and bdb
Posted by Jens Peters <jp...@gmx.de>.
Senthil Kumaran S wrote:
> If this is specific for Windows, a check for os type (os.name) would be
> good, along with a comment explaining why we have this as dglasser has
> pointed out.
>
I did a rerun on a much faster system and there I didn't got that
failure, thus I assume this is just a result of my 6 year old laptop ;).
According to Kamesh's report (
http://svn.haxx.se/dev/archive-2008-06/0151.shtml) I guess also that
this is not a windows only problem, but just occurs more often on
windows. I have attached a new patch with some explanation.
[[[
Fix a depth-tests-23 failure (Access is denied) for the combination
[bdb] x [svn | dav].
* subversion/tests/cmdline/depth_tests.py
Insert one second sleep to give slower systems the time to release
the Berkeley DB shared memory files before recreating the repository.
Patch by: jpeters7677@gmx.de
]]]
Re: [PATCH] Let depth-tests-23 pass on Windows for dav|svn and bdb
Posted by Senthil Kumaran S <se...@collab.net>.
Hi Jens,
Jens Peters wrote:
> # Our testing module
> import svntest
> @@ -1264,6 +1265,7 @@
>
> # Reset and do it again, this time from above the working copy.
> svntest.main.safe_rmtree(wc)
> + time.sleep(1)
> wc, ign_a, ign_b, ign_c = set_up_depthy_working_copies(sbox, empty=True)
> expected_output = svntest.wc.State(wc, {
If this is specific for Windows, a check for os type (os.name) would be good,
along with a comment explaining why we have this as dglasser has pointed out.
--
Senthil Kumaran S
http://www.stylesen.org/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: [PATCH] Let depth-tests-23 pass on Windows for dav|svn and bdb
Posted by David Glasser <gl...@davidglasser.net>.
I'm not sure if this is the best way to fix it, but at the very least
there should be a comment explaining the cause for the sleep.
--dave
2008/6/1 Jens Peters <jp...@gmx.de>:
> Hello there,
>
> As mentioned here earlier I see depth-tests failures on my system with the
> combination svn | dav and bdb. See
> http://svn.haxx.se/dev/archive-2008-05/0166.shtml for details. I finally
> found some time to dig into it and as a result the attached patch fixes the
> issue for me.
>
> I know that this is more an ugly workaround than a real fix. But is seems to
> work fine and at least all tests are green now (which feels actually pretty
> good, +1 to release from me :) ). Any chance that this patch or anything
> similar can be committed? Note thas this patch is made against 1.5-rc8. I
> would love to make patch against trunk, but first I would like to know if it
> will be accepted.
>
> Regards, Jens
>
> [[[
> Fix a depth-tests-23 failure (WindowsError: Access is denied) on Windows for
> the combination [svn | dav] and [bdb].
>
> * subversion/tests/cmdline/depth_tests.py
> Insert one second sleep to give Windows the time to release the Berkeley DB
> file handles before recreating the repository.
> ]]]
>
>
>
> Index: subversion/tests/cmdline/depth_tests.py
> ===================================================================
> --- subversion/tests/cmdline/depth_tests.py (revision 31532)
> +++ subversion/tests/cmdline/depth_tests.py (working copy)
> @@ -20,6 +20,7 @@
>
> # General modules
> import os
> +import time
>
> # Our testing module
> import svntest
> @@ -1264,6 +1265,7 @@
>
> # Reset and do it again, this time from above the working copy.
> svntest.main.safe_rmtree(wc)
> + time.sleep(1)
> wc, ign_a, ign_b, ign_c = set_up_depthy_working_copies(sbox, empty=True)
> expected_output = svntest.wc.State(wc, {
> 'iota' : Item(status='A '),
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
> For additional commands, e-mail: dev-help@subversion.tigris.org
>
--
David Glasser | glasser@davidglasser.net | http://www.davidglasser.net/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org