Merge Bug? no, it must be user error....I hope

[pk...@connected.com]

Guys, I recently discovered an issue with a merge between my branches.
Basically, this happened:
 
I have branches currentA, currentB, newReleaseC
I merged a revision range from currentA to currentB and found changes that
are only in newReleaseC.
 
What is even crazier is that the revision range I merged was 100:105, but
the if I do a diff on newReleaseC for that range, I do not see any changes
because the change was made in r25
 
So, operator error seems unlikely, because it would have mean both keying in
the wrong branch and the wrong range.
 
 
I don't trust the operator (especially because it was probably me).  Does
anyone one know how I might track down exactly what happened?
 
I have tried simulating this situation by doing the following
            svn update -r<bad merge revision -1> currentB
            svn merge -r100:105 currentA currentB
            svn st currentB   # check results
 
            reset environment
            svn update -r<bad merge revision -1> currentB
            svn merge -r100:105 newReleaseC currentB
            svn st currentB   # check results
 
In both cases, no change came down.
 
So, I am confused.  Anyone have any ideas on paths of investigation?   
 
Thanks for the help...
--
Peter Kahn
pkahn@connected.com
 

Re: Merge Bug? no, it must be user error....I hope

[kf...@collab.net]

I think to be able to help, we might need a totally concrete example
(the actual revisions, the actual changes, the actual URLs, etc).

-Karl

pkahn@connected.com writes:
> Guys, I recently discovered an issue with a merge between my branches.
> Basically, this happened:
>  
> I have branches currentA, currentB, newReleaseC
> I merged a revision range from currentA to currentB and found changes that
> are only in newReleaseC.
>  
> What is even crazier is that the revision range I merged was 100:105, but
> the if I do a diff on newReleaseC for that range, I do not see any changes
> because the change was made in r25
>  
> So, operator error seems unlikely, because it would have mean both keying in
> the wrong branch and the wrong range.
>  
>  
> I don't trust the operator (especially because it was probably me).  Does
> anyone one know how I might track down exactly what happened?
>  
> I have tried simulating this situation by doing the following
>             svn update -r<bad merge revision -1> currentB
>             svn merge -r100:105 currentA currentB
>             svn st currentB   # check results
>  
>             reset environment
>             svn update -r<bad merge revision -1> currentB
>             svn merge -r100:105 newReleaseC currentB
>             svn st currentB   # check results
>  
> In both cases, no change came down.
>  
> So, I am confused.  Anyone have any ideas on paths of investigation?   
>  
> Thanks for the help...
> --
> Peter Kahn
> pkahn@connected.com
>  

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org