You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by "D.J. Heap" <dj...@shadyvale.net> on 2003/07/12 08:01:50 UTC
Tests on Windows
I'm having several ra_local test failures on Windows as of r6455. I
haven't run them for a couple of weeks, so I don't know when it started.
The same revision on the same machine dual-booted to RH9 passes all
ra_local tests just fine. I've posted the failed test log to
http://www.shadyvale.net/files/tests.log if anyone cares to take a look.
Some of the tree-unequal failures look like the test is expecting the
output in a certain order, so even though the tree has the same content,
it is not coming in the order the test expects and so fails. I'm not
sure if that's really a problem or if it's legit?
And a couple of the failures look like they could be due to the
(presumed) FS race on Windows, but there used to be just one test
(commit-21, I think) that would fail now and then. Now I get between 6
and 10 failures every time.
Is anyone else running the tests regularly on Windows or seeing failures
like these or have any helpful ideas or anything...?
DJ
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Tests on Windows
Posted by Ben Collins-Sussman <su...@collab.net>.
"D.J. Heap" <dj...@shadyvale.net> writes:
> Yep, I'm on the list and have noticed that HP-UX fails every night and
> Linux passes every night...but I didn't see any other platforms.
Yes, HP-UX has been failing since the dawn of time, I don't even
notice it anymore. :-)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
RE: Tests on Windows
Posted by Sander Striker <st...@apache.org>.
> From: D.J. Heap [mailto:dj@shadyvale.net]
> Sent: Saturday, July 12, 2003 5:29 PM
> Yep, I'm on the list and have noticed that HP-UX fails every night and
> Linux passes every night...but I didn't see any other platforms.
HP-UX fails in an apr mmap cleanup operation in the diff tests.
I've been meaning to look at this with Cliff Woolley. We need to find
some sync'ed time to do so ;)
Sander
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Tests on Windows
Posted by "D.J. Heap" <dj...@shadyvale.net>.
Ben Collins-Sussman wrote:
> "D.J. Heap" <dj...@shadyvale.net> writes:
>
>
>>Some of the tree-unequal failures look like the test is expecting the
>>output in a certain order, so even though the tree has the same
>>content, it is not coming in the order the test expects and so fails.
>
>
> You misunderstand the tests. The whole reason the tests continually
> compare tress is to exactly *avoid* the issue of ordering. :-)
>
> So the errors you see aren't about ordering; it about the fact that
> the expected and actual trees have different structures. It's the
> standard test failure message. Take a read through
> subversion/tests/client/cmdline/README to understand the big picture.
>
I've read the README, but it was a while ago. I missed the real failure
description above the tree, sorry...I shouldn't have been up so late. :P
>
>
>>And a couple of the failures look like they could be due to the
>>(presumed) FS race on Windows, but there used to be just one test
>>(commit-21, I think) that would fail now and then. Now I get between
>>6 and 10 failures every time.
>
>
> Are they the same tests failing all the time? Or is it always changing?
>
There are 6ish that fail every time, and a few others that seem
randomish. I'll start tracking them closer again, but was kind of
alarmed at the sudden jump in failures from the last time I ran them --
it doesn't appear that they get run on Windows much. I'll also start
running them on another machine as well.
>
>>Is anyone else running the tests regularly on Windows or seeing
>>failures like these or have any helpful ideas or anything...?
>
>
> We have an 'svn-breakage' mailing list -- you should subscribe.
> Branko has nightly build/test boxes that mail their results to that
> list. He used to test win32 nightly, but hasn't been doing so lately.
Yep, I'm on the list and have noticed that HP-UX fails every night and
Linux passes every night...but I didn't see any other platforms.
DJ
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Tests on Windows
Posted by Ben Collins-Sussman <su...@collab.net>.
"D.J. Heap" <dj...@shadyvale.net> writes:
> I'm having several ra_local test failures on Windows as of r6455. I
> haven't run them for a couple of weeks, so I don't know when it
> started. The same revision on the same machine dual-booted to RH9
> passes all ra_local tests just fine. I've posted the failed test log
> to http://www.shadyvale.net/files/tests.log if anyone cares to take a
> look.
>
That's kind of mysterious; I know that Branko said that the 0.25
tarball passed all tests on win32.
> Some of the tree-unequal failures look like the test is expecting the
> output in a certain order, so even though the tree has the same
> content, it is not coming in the order the test expects and so fails.
You misunderstand the tests. The whole reason the tests continually
compare tress is to exactly *avoid* the issue of ordering. :-)
So the errors you see aren't about ordering; it about the fact that
the expected and actual trees have different structures. It's the
standard test failure message. Take a read through
subversion/tests/client/cmdline/README to understand the big picture.
> And a couple of the failures look like they could be due to the
> (presumed) FS race on Windows, but there used to be just one test
> (commit-21, I think) that would fail now and then. Now I get between
> 6 and 10 failures every time.
Are they the same tests failing all the time? Or is it always changing?
> Is anyone else running the tests regularly on Windows or seeing
> failures like these or have any helpful ideas or anything...?
We have an 'svn-breakage' mailing list -- you should subscribe.
Branko has nightly build/test boxes that mail their results to that
list. He used to test win32 nightly, but hasn't been doing so lately.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org