git-svn mergeinfo support performance problem

git-svn mergeinfo support performance problem

[Andrew Myrick]

I've been testing git-svn v1.6.6-rc3's mergeinfo support on a large
svn repository (60,000+ revisions, 20+ GiB) that uses a very
branch-heavy integration model in which every change gets its own
branch before being committed to trunk.  As a result of the model,
there are currently over 1000 lines in the svn:mergeinfo property on
trunk.  Unfortunately, with the current implementation, git svn fetch
takes more than a minute per revision pouring through all of those
merge tickets; obviously, this is too slow to be usable for my
repository.

Are there any ideas on how git svn fetch can be sped up when facing
hundreds of mergeinfo properties?  Alternatively, would it be possible
to add an argument that would ignore the merge info?  Or, is there any
maintenance I could perform on the svn repository that would help
reduce the amount of work that git-svn must do?  In the meantime, I'll
have to stick with git 1.6.5.*.

Regards,
Andrew

P.S.  This is my first post to the list.  My apologies if this issue
has already been discussed and I did not see it in the archives, or if
I have missed a more formal mechanism for filing bug reports.

Re: git-svn mergeinfo support performance problem

[Johan 't Hart]

Andrew Myrick schreef:
> I've been testing git-svn v1.6.6-rc3's mergeinfo support on a large
> svn repository (60,000+ revisions, 20+ GiB) that uses a very
> branch-heavy integration model in which every change gets its own
> branch before being committed to trunk.  As a result of the model,
> there are currently over 1000 lines in the svn:mergeinfo property on
> trunk.

Just wondering: Isnt this workflow stalling svn itself alot too?

And also:
Do you delete the branches after you reintegrated them? If so, I think 
its safe for you to cleanup the svn mergeinfo once in a while. That 
should not affect 'svn log -g' because the mergeinfo is still there in 
older revisions. I think svn benefits from this too...

Re: git-svn mergeinfo support performance problem

[Andrew Myrick]

On Sat, Dec 19, 2009 at 4:28 PM, Johan 't Hart <johanthart@gmail.com> wrote:
> Andrew Myrick schreef:
>>
>> I've been testing git-svn v1.6.6-rc3's mergeinfo support on a large
>> svn repository (60,000+ revisions, 20+ GiB) that uses a very
>> branch-heavy integration model in which every change gets its own
>> branch before being committed to trunk.  As a result of the model,
>> there are currently over 1000 lines in the svn:mergeinfo property on
>> trunk.
>
> Just wondering: Isnt this workflow stalling svn itself alot too?

Nope, svn seems to handle it fine.

> And also:
> Do you delete the branches after you reintegrated them? If so, I think its
> safe for you to cleanup the svn mergeinfo once in a while. That should not
> affect 'svn log -g' because the mergeinfo is still there in older revisions.
> I think svn benefits from this too...

We do not delete branches after they've been reintegrated.  Bug fix
and feature branches can get reintegrated into multiple release
branches, so it's not obvious when a branch can be deleted.  There's
already enough process overhead that it's simpler just to leave all of
the branches around.  Since this doesn't seem to affect svn's
performance, we haven't really worried about it.

-Andrew