On Thu, 2009-07-30 at 23:38 +0000, Frits Jalvingh wrote:
> Ok; got one. Bzr knew I was watching and behaved reasonably well so it took a while. The .bzr.log contains:
> Fri 2009-07-31 01:26:35 +0200
Thanks for this. Was this the commit of a merge?
Data from the callgrind says it is doing lots of head operations, and is
using the populate-from-inventory codepath, rather than the iter_changes
based code path (because we disable that for older repository formats
with multiple parents in the commit).
What revno is your repository at? (bzr revno).
You can stop doing --lsprof-file now, the data you've gathered is
excellent.
On Thu, 2009-07-30 at 23:38 +0000, Frits Jalvingh wrote:
> Ok; got one. Bzr knew I was watching and behaved reasonably well so it took a while. The .bzr.log contains:
> Fri 2009-07-31 01:26:35 +0200
Thanks for this. Was this the commit of a merge?
Data from the callgrind says it is doing lots of head operations, and is from-inventory codepath, rather than the iter_changes
using the populate-
based code path (because we disable that for older repository formats
with multiple parents in the commit).
What revno is your repository at? (bzr revno).
You can stop doing --lsprof-file now, the data you've gathered is
excellent.
-Rob