Jon Daley wrote:
> On Tue, 10 Jul 2007, Graham Bloice wrote:
>> Recently a user had a conflict after a merge into the working copy, so
>> they
>> fired up TortoiseMerge. In the options they had the radio button set to
>> "Ignore all whitespaces" and this had the effect of hiding the conflicts
>> which were partially due to whitespaces. Even though they couldn't
>> see the
>> conflicts, they still clicked the "Mark as resolved". What happened next
>> isn't clear, but somehow they managed to commit the file to the trunk
>> with
>> the conflict markers still in the file. The ancillary conflict files
>> were
>> deleted.
>>
>> Out of this arose a number of questions:
>> 1. Should the TortoiseMerge option "Ignore all whitespaces" hide
>> conflicts
>> in files?
>
>> 2. Can anyone explain why the conflict markers were not removed. My
>> best
>> guess is that the file was locked by another process, maybe an editor,
>> but
>> shouldn't TSVN have warned about this?
> When you say "mark as resolved" you are effectively telling
> subversion - all set, I know what I am doing, go away. And so, it does.
> The file was not locked by another process. The user had a file that
> had markers in it, and he said that it was what he wanted. Subversion
> did the right thing in this case. You have an argument with
> TortoiseMerge not highlighting conflicts, even though it is only due to
> whitespace. You'll notice that I ignored your first question, since
> that one is harder than the latter two, and also, I haven't seen that
> behavior, so I can't verify if that is really what happened.
>
In all the subsequent testing I've done, TortoiseMerge removed the conflict
markers from the file. However, the user in question managed to commit with
them still in place.
I am well aware that svn will allow commits with conflict markers in the
files, I'm just trying to find out how, in this particular case, the
TortoiseMerge "Mark as resolved" button didn't remove them.
>> 3. Not really a TSVN question, but how does svn determine that a
>> conflict
>> still exists, it isn't only the presence of the ancillary files, is it
>> something in the entries file?
> A conflict exists when the merge happens (doing pretty fancy magic
> code as far as I can tell, since it works so well). The file stays in
> the conflicted state until the user tells it to stop (revert or resolved)
>
Regards,
Graham Bloice
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tortoisesvn.tigris.org
For additional commands, e-mail: users-help@tortoisesvn.tigris.org
Received on Tue Jul 10 21:14:31 2007