Mattias Engdegård wrote:
> 2 apr 2013 kl. 15.51 skrev Julian Foad:
>> FAIL: upgrade_tests.py 24: wc with add-add and del-del tree conflicts
>> That's failing in local_reason_str() with kind=none and reason=deleted,
>> so at least we need that function to handle that case.
> Quel blamage! I'm really sorry - I thought I had run all the tests, but
> apparently not.
> There is no excuse for that.
No problem, I do that sometimes.
>> This indicates that perhaps the only way to get kind=none is an upgrade
>> from a previous WC format. We could handle such unusual cases by constructing a
>> response from the separate parts, and not attempt to use different messages for
>> every possible combination. The important thing is to use well written messages
>> for the common cases. The unusual or impossible cases can just have some
>> inelegant construction like we were doing before, and then we'd be sure to
>> handle any case that might arise rather than throwing an error.
> I agree. Here is a new patch which does exactly that. It re-uses the XML strings
> for the unusual case since there really is no point in translating them.
> Make the conflict message "local %s %s, incoming %s %s upon %s"
> translatable by constructing it from longer parts, each translated,
> instead of single words (some of which were not translated at all).
> * subversion/svn/cl-conflicts.c
> (svn_cl__get_human_readable_tree_conflict_description): Build string out
> of translated sub-phrases rather than from single words.
> (local_reason_str, incoming_action_str, operation_str): New.
> (map_conflict_action_human, map_conflict_reason_human,
> action_str, reason_str): Removed.
Committed in r1465299. Thanks!
Received on 2013-04-06 22:35:26 CEST