> -----Original Message-----
> From: dglasser@gmail.com [mailto:dglasser@gmail.com] On
> Behalf Of David Glasser
> Sent: Thursday, May 17, 2007 10:56 AM
> To: D.J. Heap
> Cc: David Anderson; SVN Dev
> Subject: Re: Subversion 1.4.4 tarballs up for testing/signing
>
> On 5/17/07, D.J. Heap <djheap@gmail.com> wrote:
> > I get a test hang/failure with fsfs over ra_dav -- it
> appears svnadmin
> > test 4 isn't supplying a username/password when it shoul on
> a revprop
> > propget:
>
> DJ, do you get this hang/failure on trunk as well?
I also hang on svnadmin test 4 (empty_date) with 1.4.4. The same test
passes on trunk however.
Dave's patch works to fix 1.4.4, but since test 4 is identical on 1.4.x
and trunk, so something else is going on...
...and that something is this (maybe):
On 1.4.4 the first four tests don't create
"C:\SVN\src-trunk\Release\subversion\tests\cmdline\svn-test-work\local_t
mp" but on trunk they do! On 1.4.4 and trunk the tests always use the
option:
"--config-dir"
"C:\SVN\src-trunk\Release\subversion\tests\cmdline\svn-test-work\local_t
mp\config"
But since local_tmp doesn't exist on 1.4.4 until test 5 is run, no
authentication data for jrandom is cached until then, so when we get to
test 4's propget...
CMD: svn.exe "propget" "--revprop" "-r1" "svn:date"
"svn-test-work\working_copies\svnadmin_tests-4" "--config-dir"
"C:\SVN\src-trunk\Release\subversion\tests\cmdline\svn-test-work\local_t
mp\config"
...without --username and --password specified, we are out of luck and
the test hangs waiting for us to enter a username.
But on trunk this propget works with the cached info in
"C:\SVN\src-trunk\Release\subversion\tests\cmdline\svn-test-work\local_t
mp\config\auth"
So something has changed in the test suite architecture, but I can't
figure out what.
lgo - I'm cc:ing you in the hope you might know what changed with the
test suite to cause this, if anyone knows off the top of their head it's
probably you.
I'm sorta tempted to just apply Dave's little patch to the 1.4.4 RC and
proceed with testing/signing, but that seems dubious.
Paul B.
> And if so, does the attached patch (to trunk) fix it?
> (Sorry, not in an environment where I can really build/test
> right now.)
>
> (Although, hmm, it looks to me like some other tests in that
> file on trunk should have the same problem: setrevprop and
> verify_windows_paths_in_repos; these aren't in 1.4.x though.)
And both those tests call sbox.build()
Paul B.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri May 18 00:22:09 2007