Re: [PATCH] Allow the merge tool to abort the merge

On Dec 30, 2008, at 8:49 PM, Stefan Sperling wrote:

> On Tue, Dec 30, 2008 at 10:22:12AM -0800, John Gardiner Myers wrote:
>> Stefan Sperling wrote:
>>> Can you post the script?
>> The merge tool script I've been testing with is just something I hand
>> edit to suit whatever case I'm testing at the moment. The current
>> incarnation is:
>>
>
> [snip]
>
> Thanks!
>
>> Please let me know what else I need to do in order to get any of my
>> three outstanding patches committed.
>
> I'd like to get more people involved in this discussion.
>
> I don't feel confident enough to give exit 2 a special meaning
> without having heard opinions from people who have worked on the
> external merge tool support. I've put them in Cc to make them aware
> of this. Augie, Eric, please read this thread and let us know what
> you think.

I just read through the thread, and I think that, while there's
definitely a bug in the current implementation (marking a file with
conflict markers as resolved), the proposed change doesn't feel right
- we don't really know what an exit code of 2 means to other merge
tools.

Why isn't it good enough for the operation to abort as happens now
with a non-{0,1} exit code? If we fix the bug in the existing
implementation, wouldn't that result in a working copy with conflicted
files?

Augie

>
>
>> I believe the discussion on the
>> semantics of exit 1 to largely be a digression, fixable in future
>> patches.
>
> That's true, but it's quite an important digression nonetheless. :)
>
> Thanks,
> Stefan
>
> ------------------------------------------------------
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=996515

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=996975