OLPC & package dependency growth

[seth vidal]

On Wed, 2008-07-09 at 16:34 -0400, seth vidal wrote:
> On Wed, 2008-07-09 at 15:35 -0400, Jeremy Katz wrote:
> > On Wed, 2008-07-09 at 15:18 -0400, seth vidal wrote:
> > > On Wed, 2008-07-09 at 11:44 -0400, seth vidal wrote:
> > > > Taking a list of pkgs, resolving out all of their deps then calculating
> > > > installed size (not calculating vs an existing install, but raw
> > > > installed size) shouldn't take too much code at all. I'll see what I can
> > > > hack up today.
> > > > 
> > > > It won't take into account overlaid files (like docs and manpages, nor
> > > > multilib binaries) but it should give you a pretty good upper range
> > > > (sort of).
> > > 
> > > A start of what it should do. I'll fix up the kickstart parsing, among
> > > other things, later - once I get some confirmation is even remotely in
> > > the ballpark.
> > > 
> > > http://skvidal.fedorapeople.org/misc/installed-size.py
> > 
> > The thing which becomes important to see is growth (or shrinkage) in
> > packages as well as what new packages/removed packages there are.  Which
> > involves fiddly questions of growth thresholds and human analysis of the
> > output
> 
> I can output a csv of the pkgs that would be installed.
> 

as we discussed in jabber - I've changed:
http://skvidal.fedorapeople.org/misc/installed-size.py

to output:

size\tname.arch

for all the pkgs in the requested trees. If the numbers look right-ish
I'll work on getting it to take a ks.cfg as its input.

-sv