On Fri, Mar 20, 2009 at 08:12, Daniel Shahaf <d.s_at_daniel.shahaf.name> wrote:
> Daniel Shahaf wrote on Thu, 19 Mar 2009 at 16:13 -0700:
>> +# Regression test for r36686.
>> +def exclude_keeps_hidden_entries(sbox):
>> + Â "'up --set-depth exclude' doesn't lose entries"
>> +
>> + Â sbox.build()
>> + Â wc_dir = sbox.wc_dir
>> +
>> + Â A_path = os.path.join(wc_dir, 'A')
>> + Â os.chdir(A_path)
>> +
>> + Â # the second 'up' used to cause the entry of 'C' to be lost.
>> + Â svntest.main.run_svn(None, 'up', '--set-depth', 'exclude', 'C')
>> + Â svntest.main.run_svn(None, 'up', '--set-depth', 'exclude', 'D')
>> + Â # we could grep the 'entries' file, but...
>> + Â expected_stderr = ".*svn: 'C' is already under version control.*"
>> + Â svntest.actions.run_and_verify_svn(None, None, expected_stderr,
>> + Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â 'mkdir', 'C')
>> +
>
> The reason I use 'mkdir' here is that 'info' doesn't acknowledge the
> existence-in-repos of 'C'; for example:
>
> Â Â % cd greek-tree
> Â Â % svn up --set-depth exclude A
> Â Â D Â Â Â Â A
>
> Â Â ### The next two have different outputs (good):
> Â Â % svn mkdir A
> Â Â ..\..\..\subversion\libsvn_wc\adm_ops.c:1394: (apr_err=150002)
> Â Â svn: 'A' is already under version control
>
> Â Â % svn mkdir foo
> Â Â A Â Â Â Â foo
>
> Â Â ### The next two don't (not good):
> Â Â % svn info foo
> Â Â foo: Â (Not a versioned resource)
>
> Â Â ..\..\..\subversion\svn\info-cmd.c:562: (apr_err=200000)
> Â Â svn: A problem occurred; see other errors for details
>
> Â Â % svn info A
> Â Â A: Â (Not a versioned resource)
>
> Â Â ..\..\..\subversion\svn\info-cmd.c:562: (apr_err=200000)
> Â Â svn: A problem occurred; see other errors for details
>
> Â Â %
>
>
> Shouldn't 'info' act differently for non-existent targets and for
> excluded targets?
Definitely!
Sounds like an XFAIL test in the making :-)
Cheers,
-g
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1364209
Received on 2009-03-20 13:22:05 CET