[Svnmerge] Unable to get --bidirectional flag to work.

[JL]

Hi all,

I've spend some more time on this and I still can't figure out how to 
get svnmerge.py to ignore reflected revisions.  The --bidirectional 
switch just doesn't seem to be working.

Are there any steps I can run so people can see what is occurring or 
other information I can provide?  It is this bidirectional support which 
is the main reason I wish to use the script and since it seems to work 
for everyone else, I can't see why it isn't working for me.

Thanks in advance,
Jeremy

JL wrote:
> Hi
>
> I probably didn't give enough information last time.
>
> Thanks for all your replies.  Here is a log of what I'm doing.  I 
> think I'm using it correctly, but I can't get the bidirectional option 
> to work in the way I thought it should.
>
> 1) Go to a clean checkout of the trunk and do an update.
> 2) Merge in the changes from my branch to the trunk.
> 3) Go back to my branch
> 4) Check to see what is available to be merged.  However this last 
> step is showing the commit from step 3.
>
> I'm on Windows XP.  I've tried both the python.py script and the 
> windows prebuilt version.  However python is running under cygwin.
>
> Here are the steps I'm doing.
>
> 1) Clean checkout of trunk and branch
>
> D:\Home\jeremyl\clean>cd trunk
>
> D:\Home\jeremyl\clean\trunk>svn update
> At revision 61.
>
> D:\Home\jeremyl\clean\trunk>cd ..\BR_CODELINE_phase1
>
> D:\Home\jeremyl\clean\BR_CODELINE_phase1>svn update
> A    testing
> A    testing\unit
> A    testing\unit\core
> A    testing\unit\reference
> A    testing\system
> A    testing\doc
> Updated to revision 61.
>
> 2) Merge my branch into the trunk
> D:\Home\jeremyl\clean\BR_CODELINE_phase1>cd ../trunk
>
> D:\Home\jeremyl\clean\trunk>svnmerge.py avail --bidirectional
> 55-61
>
> D:\Home\jeremyl\clean\trunk>svnmerge.py merge --bidirectional
> property 'svnmerge-integrated' set on '.'
>
> property 'svnmerge-blocked' deleted from '.'.
>
> U    prototypes\RgbBitmapEffect\readme.htm
> A    testing
> A    testing\unit
> A    testing\unit\core
> A    testing\unit\reference
> A    testing\system
> A    testing\doc
>
> property 'svnmerge-integrated' set on '.'
>
> property 'svnmerge-blocked' deleted from '.'.
>
>
> D:\Home\jeremyl\clean\trunk>svn commit . -F svnmerge-commit-messag
> e.txt
> Sending        .
> Sending        prototypes\RgbBitmapEffect\readme.htm
> Adding         testing
> Adding         testing\doc
> Adding         testing\system
> Adding         testing\unit
> Adding         testing\unit\core
> Adding         testing\unit\reference
> Transmitting file data .
> Committed revision 62.
>
> 3) Go back to my branch
> D:\Home\jeremyl\clean\trunk>cd ..\BR_CODELINE_phase1
>
> D:\Home\jeremyl\clean\BR_CODELINE_phase1>svn update
> At revision 62.
>
>
> 4) See what is available.  Revision 62 is listed as available even 
> though it is a merge from the branch.
> D:\Home\jeremyl\clean\BR_CODELINE_phase1>svnmerge.py avail 
> --bidirectional
> 20,22,51,54,62
>
> Thanks!
> Jeremy
>
> Raman Gupta wrote:
>> Dustin J. Mitchell wrote:
>>  
>>> On Wed, May 14, 2008 at 9:43 PM, JL <jeremyl at research.canon.com.au> 
>>> wrote:
>>>    
>>>> When I commit these merges into the trunk I get a revision number. 
>>>> (eg,
>>>> revision 52).   The problem is when I go back to my branch and use
>>>> svnmerge to update the branch with the latest trunk changes, it sees
>>>> revision 52 as a revision available for merging, and when it does the
>>>> merge, it conflicts on all the binary files.  How do I indicate that
>>>> revision 52 was actually a merge from the branch and therefore 
>>>> there is
>>>> no need to merge it back to the branch?  I can 'block' it however this
>>>> doesn't seem to be the right way of doing it.
>>>>       
>>> --bidirectional is supposed to fix that.  Are you always using that
>>> flag (even with the 'avail' subcommand)?
>>>     
>>
>> Or grab the latest trunk version of svnmerge.py which should ignore
>> rev 52 automatically, even without --bidirectional.
>>
>> Cheers,
>> Raman
>>
>>
>>   
>
>