Python 2.6 update with portmaster

[Wesley Shields]

On Wed, Jun 10, 2009 at 12:44:42PM -0800, Mel Flynn wrote:
> On Wednesday 10 June 2009 05:26:47 Wesley Shields wrote:
> 
> > It will be quite slow compared to using pkg_which (the normal method) so
> > be patient.
> 
> Does pkg_which use the pkgdb? If so, maybe we could have pkg_info grow a cache 
> for future encounters, or we can create one:

My guess is that pkg_which is using pkgdb, which is why it is
significantly faster.

> % grep -v '^@' /var/db/pkg/*/+CONTENTS >/tmp/pkg_info.cache
> % FOUND=`sed -ne 's,^/var/db/pkg/\(.*\)/+CONTENTS:lib/python2.6/site-
> packages/xcbgen/error.py$,\1,p' /tmp/pkg_info.cache`; echo $FOUND
> xcb-proto-1.4
> 
> So:
> 's,^${PKG_DBDIR}/\(.*\)/+CONTENTS:$$f$$,\1,p'

I think there is a SoC project in the works to make our pkg utilities
much better. I don't know any of the details but I can only hope that
speed is one of the things being improved.

> > [1]: http://people.freebsd.org/~wxs/python26-portmaster.diff
> 
> What's this for?
> +JOBS=	sysctl -n kern.smp.cpus
> 
> Should be != if it's gonna be used, but it's not referenced anywhere else	 
> that I can see.

It's used in XARGS_CMD to provide a bit of parallelism in an effort to
be a bit faster. In my opinion the number can be increased much further
than kern.smp.cpus but I'm not in a position to figure out how far to
push it. The value used is better than nothing.

-- WXS