[Nix-dev] Re: fetch<dvcs> - optimization via local mirroring.

Michael Raskin 7c6f434c at mail.ru
Fri Aug 20 01:35:55 CEST 2010


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

On 08/20/2010 03:23 AM, Yury G. Kudryashov wrote:
>> Hello.
>> With fetchgit we sometimes have a problem of having to pull entire
>> history over and over again. For fetchmtn I implemented possibility to
>> specify local database and fetch needed revisions to it first (and only
>> then to check out from it).
>> I used the Monotone convention of giving branches unique names - using
>> the reversed domain name prefixes, popularized by Java.
>> Obviously with some encoding of source URLs into branch names that
>> should be possible for some other fetchers in Nixpkgs as well. Is there
>> anybody knowing those VCSes enough to do that and willing to do that?
> I can try for fetchgit.

Thanks

> How do you deal with chroot builds? Require adding "global DB dir" to build-
> chroot-dirs?

Why build a fix-output derivation in chroot?

> BTW, current nix-prefetch-git is MUCH more clever than fetchgit. As a 
> result, they produce different results ;).

I'd call that a bug..
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.15 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iQEcBAEBAgAGBQJMbb/aAAoJEE6tnN0aWvw3CTQH+gMhykPcw7m7g7AIZFNibAlc
a6XfPXkOklqdw5AS1W2w6jo+8K9BQRXXb+uJ2YviwaAVz8m2SfbRclvJwp3CeSvM
CTKIT2iVvRg6DU+uG9a8biLwcAtIn3CMzMb1mc0lNMSlDUkdCaI3GzdLhp0Y4pLm
521ThCL3E9AYopQyBj24Jo4fQ+DUrb5yPnz/8XE3SoIkSnUta+6ZOdTNERDYqGVj
Qmkv8U9OR4KJx98OEU1ABGoQeusPpsymY4VRfQiKd+aIS+Oyyo6sQEevmZyfYePJ
lsmJfEuEMacQim9qXTLYssVnD4u+Li4WvJYFWnxOXBk/cvadix7Osl1N6PJxVwo=
=P558
-----END PGP SIGNATURE-----



More information about the nix-dev mailing list