Mark Phippard wrote:
> On 8/20/07, Hyrum K. Wright <hyrum_wright@mail.utexas.edu> wrote:
>> Mark Phippard wrote:
>>> Using my sample merge tracking repository that is used in the svn log -g tests:
>>>
>>> $ svn blame -g file:///Users/mphippard/repositories/merge-tracking/trunk/index.html
>>>
>>> Works OK.
>>>
>>> $ svn blame -g file:///Users/mphippard/repositories/merge-tracking/trunk/products/index.html
>>> svn: Can't open file '/tmp/tmp.2.tmp': No such file or directory
>> Fixed in r26207.
>
> It now runs to completion, but I see user "merger" from r12 in the
> output which implies there is a bug of some sort. I think this was a
> case where branchA was merged to branchB in r12 and then to trunk
> later. So maybe it did not follow the merge back to its source?
Oh, good point. Instead of just two possible states a particular
revision can be in (merged and non-merged), we really have three
possible states. The revision could be a merged revision with content
changes, a merged revision which is also a merging revision, and a
non-merged/non-merging revision.
If the revision is a merging revision, we don't want it to show up as a
revision which changed content, even if it is also a merging revision.
I'll need to think about this for a minute...
>>> $ svn blame -g file:///Users/mphippard/repositories/merge-tracking/trunk/products/roadmap.html
>>> Bus error
>> Fixed in r26211.
>
> This still crashes for me on OSX. I have not tried it on Windows.
Hmm, I don't have any problems on Linux. Do you know where it crashes?
-Hyrum
Received on Tue Aug 21 16:39:51 2007