slower performance of log and annotate in 1.6?

[John Arbash Meinel]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Martin Pool wrote:
> The benchmarks <http://benchmark.bazaar-vcs.org/usertest/log/usertest.log> running on Orcadas seem to show log and annotate may be slower in trunk than they were in 1.5, which would be a concern for 1.6.  It may be that this is just measurement error on that machine, so there is no immediate need to panic, but if anyone either knows of a change that may have caused this or would like to try to reproduce it that would be helpful.
> 
> 
> 1:min bzr.dev	bzr.1.3	bzr.1.5	Action
> 1.23	49.437	44.208	40.159	script_common.LogTask:log
> 1.84	77.094	41.851	48.409	script_common.LogTask:logOneFile
> 1.20	8.787	8.733	7.332	script_common.AnnotateTask:annotate
> 
> 1.20	7.627	7.837	6.356	script_common.TagsTask:tags
> 
> 
> -- 
> Martin <http://launchpad.net/~mbp/>
> 

Well, I notice that 'bzr annotate' is the same time between 1.3 and bzr.dev.

If I was guessing, I would say that we made access to the per-file graph
slower than it used to be. This would impact logOneFile and annotate.
But I'm purely speculating at this point.

I know for 'annotate', Robert also changed it to annotate the working
tree, which would be a t least 1 more 'diff' operation. Though that
shouldn't account for 1.5s of time.

John
=:->
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (Cygwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAkignTgACgkQJdeBCYSNAAM0fgCeLy9MqSsjxZrSFw+NsIDjaiJA
SrYAn1nc2jh9uU4DCbGvjOfbp/V0edof
=Xh/u
-----END PGP SIGNATURE-----