Re: Trimming gdk-pixbuf repository

[Yasushi SHOJI <yashi atmark-techno com>]

At Sat, 21 May 2011 03:37:38 +0200,
Olav Vitters wrote:
> 
> On Thu, Mar 17, 2011 at 09:06:32PM +0000, Alberto Ruiz wrote:
> > I am looking for comments from git experts on any of the operations
> > I'm doing and how feasible would be to rewrite the git.gnome.org
> > repository. The resulting repository can be found in my github
> > site[0].
> 
> Any git experts who can look at this again?
> 
> https://mail.gnome.org/archives/gtk-devel-list/2011-March/msg00066.html

I'm not a git expert per se, but I would be against the idea if I was
asked for the following reasons:

 - the trim changes the SHA1 ID
   check the commit with PIXBUF_0_0 tag; the cuurent gdk-pixbuf repo
   on git.gnome.org has f0723b00 but the trimed one has 534df8c.  this
   is due to git's _feature_.

 - we usually clone a few times
   most of us do not re-clone every time we hack on it.  rather, we
   keep the cloned repo handy and reuse all the time.

 - there is another way to reduce the transfer size (if you have gtk+ repo)
   you can tell git to reference gtk+ git repo when you clone gdk-pixbuf

	$ /usr/bin/time git clone --reference gtk+ git://git.gnome.org/gdk-pixbuf
	Cloning into gdk-pixbuf...
	remote: Counting objects: 30651, done.
	remote: Compressing objects: 100% (7683/7683), done.
	remote: Total 30184 (delta 25544), reused 26364 (delta 22362)
	Receiving objects: 100% (30184/30184), 36.57 MiB | 375 KiB/s, done.
	Resolving deltas: 100% (25544/25544), completed with 417 local objects.
	18.33user 1.47system 2:31.32elapsed 13%CPU (0avgtext+0avgdata 430768maxresident)k
	541864inputs+210072outputs (102major+51069minor)pagefaults 0swaps

   it is bigger than trimed one (29.40 MiB) but 6 MiB is negligible
   IMO.  here is the log of full clone via my slow wireless
   connection, just in case.

	$ /usr/bin/time git clone git://git.gnome.org/gdk-pixbuf
	Cloning into gdk-pixbuf...
	remote: Counting objects: 229825, done.
	remote: Compressing objects: 100% (33082/33082), done.
	remote: Total 229825 (delta 197326), reused 228509 (delta 196355)
	Receiving objects: 100% (229825/229825), 133.96 MiB | 413 KiB/s, done.
	Resolving deltas: 100% (197326/197326), done.
	55.00user 3.84system 5:41.92elapsed 17%CPU (0avgtext+0avgdata 110496maxresident)k
	12872inputs+575600outputs (10major+15953minor)pagefaults 0swaps

just my two cents.
-- 
          yashi