| 1 |
On Mon, 17 Dec 2007 18:46:12 -0700 |
| 2 |
Joe Peterson <lavajoe@g.o> wrote: |
| 3 |
|
| 4 |
> What about storing a copy of the EAPI in the Manifest file - when |
| 5 |
> "ebuild ... digest" is done? That way, it will always match the one |
| 6 |
> authoritative "post-source" EAPI setting, since changing the ebuild |
| 7 |
> will require a new digest run anyway. We have control over the |
| 8 |
> format of Manifest, so parsing it can be fast. |
| 9 |
|
| 10 |
No chance. |
| 11 |
|
| 12 |
> If Manifest is not a good candidate, put them in an optional "EAPI" |
| 13 |
> file (which can then also be included in the Manifest). If the |
| 14 |
> external EAPI info for an ebuild is not found, then drop back to |
| 15 |
> assuming it does not have a defined pre-source EAPI. |
| 16 |
|
| 17 |
While I also dislike using the filename extension this would be an even |
| 18 |
greater mess (we're trying to reduce the number of files in the |
| 19 |
repository, so adding another 20k tiny files would require an |
| 20 |
extremely good reason). |
| 21 |
|
| 22 |
My personal opinion is that EAPI is the wrong angle to solve the |
| 23 |
issue with versioning, that's something that should be dealt with at |
| 24 |
the repository level (as it will eventually impact comparison rules |
| 25 |
between "old-style" versions as well, and those by definition can't be |
| 26 |
changed on a per-package base). |
| 27 |
That leaves the argument about global-scope functions, for which there |
| 28 |
are workarounds (not solutions), and without actual use cases it's a |
| 29 |
bit hard to evaluate cost and benefits. |
| 30 |
|
| 31 |
Marius |
| 32 |
-- |
| 33 |
gentoo-dev@g.o mailing list |
Archives