[RFC] temporarily disable format-display after creating repositories/branches?
robert.collins at canonical.com
Thu Feb 19 21:44:00 GMT 2009
On Thu, 2009-02-19 at 08:55 -0600, John Arbash Meinel wrote:
> Is "bzr init-repo bzr+ssh://" something we are really trying to
> optimize? I can see optimizing "bzr push bzr+ssh://" which creates a
> repository, but the former is not really a performance critical piece
> As for me, I've come to like the display, so I'd rather not see it go.
> (though I'll admit I don't like it enough to spend the time to bring
> back if it was voted off).
Its been beaten into a shape where I'm reasonably happy with it to; this
isn't an attempt to get rid of it the back-door-way.
> So IMO, I'm willing to accept the extra 11 round trips for a
> non-critical command, rather than either asking anyone to spend a lot
> time optimizing it, or getting rid of a nice UI feature.
11 round trips is a lot of waiting time - for me its about 4 seconds. So
with the major causes of slowness removed I fully expect to see bug
reports at some point that say 'I just noticed, bzr -q init-repo bzr
+ssh://foo is twice as fast as it is without the -q'.
> If you feel strongly, you can get rid of it, but it wouldn't be my
I think it makes a bad impression for folk looking to benchmark bzr to
be slower than we need to be, for commands that they expect to be fast.
cmd_init also does this but I don't have figures for it yet. When we get
branch creation on push sorted out I'll probably be adding an rpc
tracking test there, and I'll see how it goes.
I don't feel very strongly, or I would have put up a [MERGE] rather than
discussing it :)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 197 bytes
Desc: This is a digitally signed message part
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20090220/8aa8d1d4/attachment.pgp
More information about the bazaar