On Saturday, February 1, 2003, at 08:21 AM, Philip Martin wrote:
> rooneg@tigris.org writes:
>
>> Author: rooneg
>> Date: 2003-01-30 20:16:39 -0600 (Thu, 30 Jan 2003)
>> New Revision: 4683
>>
>> Modified:
>> trunk/subversion/libsvn_client/diff.c
>> Log:
>> * subversion/libsvn_client/diff.c
>> (do_single_file_merge): detect the text state before we call the
>> notify
>> function.
>>
>> Modified: trunk/subversion/libsvn_client/diff.c
>> ======================================================================
>> ========
>> --- trunk/subversion/libsvn_client/diff.c (original)
>> +++ trunk/subversion/libsvn_client/diff.c Thu Jan 30 20:16:44 2003
> [...]
>> @@ -923,6 +927,15 @@
>> &merge_outcome, pool));
>> if (merge_outcome == svn_wc_merge_conflict)
>> text_state = svn_wc_notify_state_conflicted;
>> + else if (merge_outcome == svn_wc_merge_merged)
>> + {
>> + if (already_modified)
>> + text_state = svn_wc_notify_state_merged;
>> + else
>> + text_state = svn_wc_notify_state_changed;
>> + }
>> + else
>> + text_state = svn_wc_notify_state_unchanged;
>
> That's not the same as the code in merge_file_changed, although it may
> well produce the same result. It's silly to have this sort of
> duplication (or should that be mutation?), do_single_file_merge should
> be invoking the merge svn_wc_diff_callbacks_t callbacks to ensure
> consistent behaviour.
that makes sense. i'll take a closer look at this later today and see
what i can do about it.
> I note that you appear to have changed the feedback without breaking
> the regression tests, that must mean there is no test for it :(
perhaps if i'm extra motivated i'll add a new test ;-)
-garrett
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Feb 1 15:17:09 2003