On Fri, Mar 13, 2009 at 09:58, Daniel Rall <dlr_at_finemaltcoding.com> wrote:
> On Fri, Mar 13, 2009 at 1:53 AM, Daniel Rall <dlr_at_finemaltcoding.com> wrote:
> ...
>> --- trunk/subversion/tests/libsvn_wc/db-test.c Fri Mar 13 00:43:51 2009 (r36523)
>> +++ trunk/subversion/tests/libsvn_wc/db-test.c Fri Mar 13 01:53:34 2009 (r36524)
> ...
>> SVN_ERR(svn_wc__db_base_get_props(&props, db, path,
>> scratch_pool, scratch_pool));
>> + switch (status)
>> + {
>> + case svn_wc__db_status_absent:
>> + case svn_wc__db_status_excluded:
>> + case svn_wc__db_status_incomplete:
>> + case svn_wc__db_status_not_present:
>> + /* Our tests aren't setting properties on these node types, so
>> + short-circuit examination of name/value pairs. */
>> + return SVN_NO_ERROR;
>> +
>> + default:
>> + SVN_ERR_ASSERT(props != NULL);
>> + }
> ...
>
> Will all of these states always be absent of properties? If not, we
> should alter the data seeded by test_inserting_nodes(). It currently
> does not seed any property values for these cases because the
> corresponding APIs used to inject the data lack a "props" parameter.
Those states will never have properties. They're different ways to say
"the node isn't present, and <this> is why."
Thanks for the test update!!
Cheers,
-g
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1316711
Received on 2009-03-13 10:18:35 CET