You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Paul Burba <pb...@collab.net> on 2008/01/18 15:14:55 UTC
RE: svn commit: r28949 - branches/reintegrate/subversion/tests/cmdline
> -----Original Message-----
> From: kfogel@tigris.org [mailto:kfogel@tigris.org]
> Sent: Thursday, January 17, 2008 10:12 PM
> To: svn@subversion.tigris.org
> Subject: svn commit: r28949 -
> branches/reintegrate/subversion/tests/cmdline
>
> Author: kfogel
> Date: Thu Jan 17 19:12:00 2008
> New Revision: 28949
>
> Log:
> * subversion/tests/cmdline/merge_tests.py
> (basic_reintegrate): Insert an update, so this test passes again,
> but include a comment questioning the sanity of the situation.
Hi Karl,
(Ignoring the question of r28942's validity for the moment, we can
discuss in a different thread or IRC)
This test really passes for you? It fails for me (see attached log). I
think the update needs to come before the copy so 'A_COPY/D/G/tauprime'
can inherit mergeinfo from 'A_COPY'.
Paul
> Modified:
> branches/reintegrate/subversion/tests/cmdline/merge_tests.py
>
> Modified: branches/reintegrate/subversion/tests/cmdline/merge_tests.py
> URL:
> http://svn.collab.net/viewvc/svn/branches/reintegrate/subversi
> on/tests/cmdline/merge_tests.py?pathrev=28949&r1=28948&r2=28949
> ==============================================================
> ================
> ---
> branches/reintegrate/subversion/tests/cmdline/merge_tests.py
> (original)
> +++
> branches/reintegrate/subversion/tests/cmdline/merge_tests.py
> Thu Jan 17 19:12:00 2008
> @@ -9530,6 +9530,7 @@
> svntest.actions.run_and_verify_commit(wc_dir, expected_output,
> expected_status,
> None, wc_dir)
>
> +
> # Make another change on the branch: copy tau to tauprime. Commit
> # in r9.
> svntest.actions.run_and_verify_svn(None, None, [], 'cp',
> @@ -9537,12 +9538,58 @@
> 'tau'),
> os.path.join(wc_dir,
> 'A_COPY', 'D', 'G',
> 'tauprime'))
> +
> + # Update the wcs again.
> + #
> + # Note: this update had to be added because of r28942
> (which was #
> + merged into the reintegrate branch in r28947). Without this #
> + update, the update to r9 later will fail in expected disk tree:
> + #
> + # =============================================================
> + # Expected 'tauprime' and actual 'tauprime' in disk tree
> are different!
> + # =============================================================
> + # EXPECTED NODE TO BE:
> + # =============================================================
> + # * Node name: tauprime
> + # Path: A_COPY/D/G/tauprime
> + # Contents: This is the file 'tau'.
> + #
> + # Properties: {'svn:mergeinfo': '/A/D/G/tau:2-7'}
> + # Attributes: {}
> + # Children: N/A (node is a file)
> + # =============================================================
> + # ACTUAL NODE FOUND:
> + # =============================================================
> + # * Node name: tauprime
> + # Path: G/tauprime
> + # Contents: This is the file 'tau'.
> + #
> + # Properties: {'svn:mergeinfo': ''}
> + # Attributes: {}
> + # Children: N/A (node is a file)
> + #
> + # What is truly strange (IMHO) is that you can put this update #
> + *after* the 'svn cp' below, and the test still passes; but
> if # you
> + put it after the commit, the test fails with the above error.
> + # It is as if the mergeinfo inheritance happens in the
> commit step #
> + (I would have expected it to happen in the cp step).
> + #
> + # By the way, the test passes -- as one would expect -- if this #
> + update is put right before the 'svn cp tau tauprime' command #
> + immediately above.
> + expected_output = wc.State(wc_dir, {})
> + expected_status.tweak(wc_rev='8')
> + svntest.actions.run_and_verify_update(wc_dir, expected_output,
> + expected_disk,
> expected_status,
> + None, None, None,
> None, None,
> + True)
> +
> expected_output = wc.State(wc_dir, {
> 'A_COPY/D/G/tauprime' : Item(verb='Adding')
> })
> expected_status.add({'A_COPY/D/G/tauprime': Item(status='
> ', wc_rev=9)})
> svntest.actions.run_and_verify_commit(wc_dir, expected_output,
> expected_status,
> None, wc_dir)
> +
> expected_disk.add({
> 'A_COPY/D/G/tauprime' : Item(props={SVN_PROP_MERGE_INFO:
> '/A/D/G/tau:2-7'},
> contents="This is the file
> 'tau'.\n")
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: svn-unsubscribe@subversion.tigris.org
> For additional commands, e-mail: svn-help@subversion.tigris.org
>
>
Re: svn commit: r28949 - branches/reintegrate/subversion/tests/cmdline
Posted by Karl Fogel <kf...@red-bean.com>.
"Paul Burba" <pb...@collab.net> writes:
> I rebuilt and tested this morning, and the test passed. I was
> completely perplexed until I saw r28964 and the log update to r28949.
> So we are all set on this particular question yes?
Yes. (There may still be questions about r28942 itself, but the
question of how the tests should look right now is settled.)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: svn commit: r28949 - branches/reintegrate/subversion/tests/cmdline
Posted by Karl Fogel <kf...@red-bean.com>.
"Paul Burba" <pb...@collab.net> writes:
> I rebuilt and tested this morning, and the test passed. I was
> completely perplexed until I saw r28964 and the log update to r28949.
> So we are all set on this particular question yes?
Yes. (There may still be questions about r28942 itself, but the
question of how the tests should look right now is settled.)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
RE: svn commit: r28949 - branches/reintegrate/subversion/tests/cmdline
Posted by Paul Burba <pb...@collab.net>.
> -----Original Message-----
> From: Karl Fogel [mailto:kfogel@red-bean.com]
> Sent: Friday, January 18, 2008 8:01 PM
> To: Paul Burba
> Cc: dev@subversion.tigris.org
> Subject: Re: svn commit: r28949 -
> branches/reintegrate/subversion/tests/cmdline
>
> "Paul Burba" <pb...@collab.net> writes:
> > (Ignoring the question of r28942's validity for the moment, we can
> > discuss in a different thread or IRC)
> >
> > This test really passes for you? It fails for me (see
> attached log).
> > I think the update needs to come before the copy so
> 'A_COPY/D/G/tauprime'
> > can inherit mergeinfo from 'A_COPY'.
>
> Yes, it really passes for me (over ra_local, same as you
> ran), and it fails without that change. I checked this
> several times, to make sure. (Furthermore, when it failed
> for me, the error was not "Couldn't find node 'tauprime' in
> expected disk tree", as it was for
> you.)
>
> There is some seriously randomizing force at work here... :-(
I rebuilt and tested this morning, and the test passed. I was
completely perplexed until I saw r28964 and the log update to r28949.
So we are all set on this particular question yes?
> I wasn't quite sure of the motivations behind r28942. It
> seemed like it would prohibit a lot of mergeinfo inheritance
> that we actually want... I'll start a separate thread on
> that, based on the original commit mail if I still have it.
>
> -Karl
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: svn commit: r28949 - branches/reintegrate/subversion/tests/cmdline
Posted by Karl Fogel <kf...@red-bean.com>.
"Paul Burba" <pb...@collab.net> writes:
> (Ignoring the question of r28942's validity for the moment, we can
> discuss in a different thread or IRC)
>
> This test really passes for you? It fails for me (see attached log). I
> think the update needs to come before the copy so 'A_COPY/D/G/tauprime'
> can inherit mergeinfo from 'A_COPY'.
Yes, it really passes for me (over ra_local, same as you ran), and it
fails without that change. I checked this several times, to make
sure. (Furthermore, when it failed for me, the error was not
"Couldn't find node 'tauprime' in expected disk tree", as it was for
you.)
There is some seriously randomizing force at work here... :-(
I wasn't quite sure of the motivations behind r28942. It seemed like
it would prohibit a lot of mergeinfo inheritance that we actually
want... I'll start a separate thread on that, based on the original
commit mail if I still have it.
-Karl
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org