Re: [PATCH/RFC/GSoC 01/17] perf: introduce performance tests for git-rebase

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Thomas,

On Fri, 18 Mar 2016, Thomas Gummerer wrote:

> On 03/16, Johannes Schindelin wrote:
> > 
> > On Wed, 16 Mar 2016, Paul Tan wrote:
> > 
> > > On Wed, Mar 16, 2016 at 3:58 PM, Johannes Schindelin
> > > <Johannes.Schindelin@xxxxxx> wrote:
> > > >
> > > > On Sat, 12 Mar 2016, Paul Tan wrote:
> > > >
> > > >> diff --git a/t/perf/p3404-rebase-interactive.sh b/t/perf/p3404-rebase-interactive.sh
> > > >> new file mode 100755
> > > >> index 0000000..aaca105
> > > >> --- /dev/null
> > > >> +++ b/t/perf/p3404-rebase-interactive.sh
> > > >> @@ -0,0 +1,26 @@
> > > >>
> > > >> [...]
> > > >>
> > > >> +test_perf 'rebase -i --onto master^' '
> > > >> +     git checkout perf-topic-branch &&
> > > >> +     git reset --hard perf-topic-branch-initial &&
> > > >> +     GIT_SEQUENCE_EDITOR=: git rebase -i --onto master^ master
> > > >> +'
> > > >
> > > > This measures the performance of checkout && reset && rebase -i. Maybe we
> > > > should only test rebase -i?
> > > 
> > > test_perf runs the same script multiple times, so we need to reset
> > > --hard at least to undo the changes of the rebase.
> > > 
> > > I think we can remove the reset if we use rebase -f and rebase onto
> > > the same base, but -f was not implemented in this patch series.
> > 
> > Hrm. rebase -f just makes the reset an implicit part of the rebase, so it
> > seems we cannot perf *just* the rebase. We are stuck with perf'ing also
> > the reset. Sad.
> 
> I had the same problem back when I was working on index-v5 and posted
> a patch series.  The discussion about it is at [1].  Maybe it could be
> worth resurrecting?
> 
> [1] http://thread.gmane.org/gmane.comp.version-control.git/1379419842-32627-1-git-send-email-t.gummerer@xxxxxxxxx

Yes, I agree that something like that is needed. The proposed commit
message suggests that things get simpler, though, while I would contend
that timings get more accurate.

And I think you could simply move the test_start command, but that's just
from a *very* cursory reading of the patch.

Ciao,
Dscho
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html



[Index of Archives]     [Linux Kernel Development]     [Gcc Help]     [IETF Annouce]     [DCCP]     [Netdev]     [Networking]     [Security]     [V4L]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux SCSI]     [Fedora Users]