| 1 |
-----BEGIN PGP SIGNED MESSAGE----- |
| 2 |
Hash: SHA1 |
| 3 |
|
| 4 |
Ciaran McCreesh wrote: |
| 5 |
> On Sat, 07 Feb 2009 15:23:18 -0800 |
| 6 |
> Zac Medico <zmedico@g.o> wrote: |
| 7 |
>>> Well, usually you don't keep intermediate or generated files in a |
| 8 |
>>> VCS, so why the metadata? |
| 9 |
>> People who distribute overlays commonly ask if it's possible to |
| 10 |
>> distribute metadata cache with the overlay. Using a format that |
| 11 |
>> doesn't rely on timestamps will allow them to distribute metadata |
| 12 |
>> cache using their existing infrastructure, which is typically git or |
| 13 |
>> subversion. In addition to overlays, it would also be useful for |
| 14 |
>> forks of the entire gentoo tree, such as the funtoo tree [1]. |
| 15 |
> |
| 16 |
> Are these people really all going to remember to run some command at |
| 17 |
> the top level of the repository before every commit, and to git add the |
| 18 |
> relevant files for everything (thus making really messy commits)? |
| 19 |
|
| 20 |
The cache can be incrementally updated by a tool such as repoman, or |
| 21 |
it can be updated in periodic batches by a cron job. The periodic |
| 22 |
batch approach may be more convenient for eclass changes affect |
| 23 |
large numbers of ebuilds. |
| 24 |
|
| 25 |
> Sticking metadata cache files under version control really is a perfect |
| 26 |
> example of doing it wrong... |
| 27 |
|
| 28 |
Well, if you want to use timestamps, the alternative is to |
| 29 |
distributors to use a protocol which preserves timestamps. This |
| 30 |
creates an unnecessary burden. Allowing distribution of metadata |
| 31 |
cache via version control systems is more flexible. |
| 32 |
- -- |
| 33 |
Thanks, |
| 34 |
Zac |
| 35 |
-----BEGIN PGP SIGNATURE----- |
| 36 |
Version: GnuPG v2.0.9 (GNU/Linux) |
| 37 |
|
| 38 |
iEYEARECAAYFAkmPX/QACgkQ/ejvha5XGaMPewCeOZYsNt2bv+CbOV58aV7isq4f |
| 39 |
wCAAnA/10jcuad5NrP3BxyFZAYWH07ot |
| 40 |
=iRw3 |
| 41 |
-----END PGP SIGNATURE----- |
Archives