Karl Fogel wrote:
> mstrap@tigris.org writes:
>> SVN r27021 from trunk. With the attached test repository, I have
>> performed various merges from branches to other branches and finally
>> back to the trunk and now get following revisions for "svn log -g
>> svn://localhost":
>>
>> r36
>> r35 result of merge from r36
>> r34 result of merge from r36, r35
>> r33 result of merge from r36, r35, r34
>> r25 result of merge from r36, r35
>> r35
>> r34 result of merge from r35
>> r33 result of merge from r35, r34
>> r34
>> r33 result of merge from r34
>> r33
>> r32
>> ...
>>
>> What is inconsistent here it that r25 is reported as result of merge
>> from r35 (as part of r36 report), but it's not reported anymore for
>> r35 itself. I'm not sure whether r25 should actually be reported
>> because I have played with svn:mergeinfo at r35 and at r36, but I
>> would anyway not expect this mixture.
>
> I'd like to understand this, but I'm embarrassed to say that I cannot
> figure out what the log -g output means. Presumably r25 cannot be the
> commit that resulted from merging the r35 changes, since r35 postdates
> r25.
>
> When I see this, for example:
>
> ----------------------------------------------------------------------
> r43 | strapetz | 2007-10-02 20:48:31 -0700 (Tue, 02 Oct 2007) | 1 line
> Changed paths:
> M /merge/branches/b21/file1.txt
> Result of a merge from: r45
>
> Changed file1.txt in b21.
> ----------------------------------------------------------------------
>
> I wonder how anything that happens in r43 can be the "result of a
> merge from r45"?
>
> I believe this is the second time this thread has come up, and that
> when I asked these questions last time, we agreed on a new phrasing
> that would be clearer, but then I never got around to making the
> change. This time, once we agree on what the output means, I will
> make the change right away!
The meaning of "Result of merge from: rXXX" is "the reason we are
showing this revision is because it was merged to the line of history of
interest in rXXX". I'm not married to that verbiage; it was more of a
placeholder than anything else. If the current wording is unclear (and
I can see how it would be), let's get it changed.
-Hyrum
Received on Thu Nov 15 19:41:41 2007