On Thu, Jan 10, 2013 at 2:15 PM, Bert Huijben <bert_at_qqmail.nl> wrote:
>
>> -----Original Message-----
>> From: Paul Burba [mailto:ptburba_at_gmail.com]
>> Sent: donderdag 10 januari 2013 19:59
>> To: Bert Huijben
>> Cc: dev_at_subversion.apache.org
>> Subject: Re: svn commit: r1431114 -
>> /subversion/trunk/subversion/svn/merge-cmd.c
>>
>> On Thu, Jan 10, 2013 at 12:38 PM, Bert Huijben <bert_at_qqmail.nl> wrote:
>> >
>> >
>> >> -----Original Message-----
>> >> From: pburba_at_apache.org [mailto:pburba_at_apache.org]
>> >> Sent: woensdag 9 januari 2013 23:04
>> >> To: commits_at_subversion.apache.org
>> >> Subject: svn commit: r1431114 -
>> /subversion/trunk/subversion/svn/merge-
>> >> cmd.c
>> >>
>> >> Author: pburba
>> >> Date: Wed Jan 9 22:04:24 2013
>> >> New Revision: 1431114
>> >>
>> >> URL: http://svn.apache.org/viewvc?rev=1431114&view=rev
>> >> Log:
>> >> Fix issue #4139 'Subversion cannot perform merge if there's a file with
>> >> the same name as directory'.
>> >>
>> >> * subversion/svn/merge-cmd.c
>> >> (svn_cl__merge): If the basename of the source is the same as the
>> >> basename of the current working directory, then assume the cwd is the
>> >> target.
>> >
>> > I never heard of and/or noticed this behavior
>>
>> Hi Bert,
>>
>> Which behavior are you referring to: The old behavior, the bug with
>> the old behavior, or the new behavior?
>>
>> The old behavior was this:
>>
>> 'svn merge ^/src/base-name .' and 'svn merge ^src/base-name' both used
>> the cwd at the merge target *unless* there is a file in the cwd with
>> the same name as the source basename. In that case the file was the
>> merge target.
>
> The question I asked was, shouldn't we make that first case think the "." is an explicit target?
>
> "I tell svn where to merge to and I would have guessed that it just did what I asked."
Ok, seems you are saying what I said before:
> Merge target: "." --> No guessing, the target is *always* the cwd.
>
> Merge target: "" --> Target guessing heuristics apply.
I'm good with that. If nobody objects I can make that change.
> That we want to be smart for the second case is a different question, but my preference would be that this explicit '.' case would be handled as the plain api.
Mildly puzzled: Which API do you mean? Everything we are talking
about is in svn_cl__merge(), prior to calling any public APIs. I
think I know what you mean...but I'm not entirely sure :-)
> (Our testsuite doesn't care: it only tests the second variant)
>
> Bert
>
--
Paul T. Burba
CollabNet, Inc. -- www.collab.net -- Enterprise Cloud Development
Skype: ptburba
Received on 2013-01-10 20:27:01 CET