On Mon, 22 Apr 2013 22:46:14 +0000 (UTC)
Duncan <1i5t5.duncan@cox.net> wrote:

> Alexis Ballier posted on Mon, 22 Apr 2013 15:40:33 +0200 as excerpted:
> 
> > "Having functional changes mixed with whitespace/cosmetics in a
> > single commit makes it hard to read and understand."
> > 
> >> Also, in this case I see only one extra hunk.
> >> But once we have proper tools (like git) we can revisit this.
> > 
> > I don't see how git helps. You'll have to commit twice then push, vs
> > commit twice with cvs.
> 
> But git commits are quite lightweight, while as someone already
> pointed out, cvs commits, if done properly with repoman, are anything
> but.
> 
> So at least in the sense that it'll be less hassle, two git commits 
> followed by a push should be much easier than two repoman and cvs
> commits.
> 

Maybe the question is rather why `repoman` takes 15 seconds on a quite
fast system in a package folder that contains 2 ebuilds and 1 metadata.

See the call graph for repoman at http://i.imgur.com/OQTUBdR.png.

A third of the time, ~5 seconds, are spent on 60,000 - 70,000 calls to
the function that is listed in the bottom; but the ebuilds barely
list any dependencies. Why are there so much calls to this?

A third of the time, ~5 seconds, are spent on visibility; if we look at
the `emerge -puDN @world` call graph (http://i.imgur.com/A93CdNR.png)
we also see visibility checks taking a prominent enough place. If this
is part is really needed, we may look at how we can improve this and at
the same time improvement the dependency calculation and repoman times.

Last third of the time, ~5 seconds, are spent on regular expressions
and mostly configuration; why is configuration so invasive here?

My thoughts are that we may be able to shave ten seconds of this time.

-- 
With kind regards,

Tom Wijsman (TomWij)
Gentoo Developer

E-mail address  : TomWij@gentoo.org
GPG Public Key  : 6D34E57D
GPG Fingerprint : C165 AF18 AB4C 400B C3D2  ABF0 95B2 1FCD 6D34 E57D