[Gnu-arch-users] Re: Working out a branching scheme

[Neil Stevens]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thursday April 01, 2004 3:42 pm, Dustin Sallings wrote:
> On Apr 1, 2004, at 12:02, Aaron Bentley wrote:
> > foo--devel--0 ----------------------------------------------->
> >
> >                foo--release--0.7------     foo--release--0.8
> >
> >                   0.7.1  0.7.2      0.7.3       0.8.1
>
>       This model makes a lot of sense to me, but you end up with a really
> large number of patches in a branch, which arch doesn't seem to handle
> very efficiently.

Tagging continuation branches doesn't reduce the number of patches 
involved, since you have to pull them all to create a tree no matter how 
many branches are involved.

I work with arch the same way I used to work with CVS: I have one mainline 
branch, and then tag stable branches from there.  It's way better than CVS 
because of the ease of backporting fixes to the stable branches.  And yes, 
this means my main branch will end up getting lots of patches.

One of my mainlines through two archives (I created a new archive to rename 
to my new domain and to get tla 1.2 signing) has over 180 patches in it 
(130 on one side, 50 on the other).  So I'm really interested in knowing:  
how many patches counts as a "large number" and what parts of arch have 
problems with that large number?

I know it IS slow to do fresh gets when there are many patches involved, 
but that's what tla cacherev is for.  I do that on all the mainline 
branches I tag from.

thanks,

- -- 
Neil Stevens - address@hidden

"It's snowing, it's snowing! God, I hate this weather."
    They Might Be Giants, __New York City__
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (FreeBSD)

iD8DBQFAbLQ7f7mnligQOmERArGRAJ94T/EwEVugxwVV5qG0s/4o0XSc+wCgglW1
aqqcgw98J3VLAkmppmZlA9Y=
=7qZC
-----END PGP SIGNATURE-----