| 1 |
On 03/21/2017 10:17 AM, Alexis Ballier wrote: |
| 2 |
> On Tue, 21 Mar 2017 09:43:37 +0100 |
| 3 |
> Kristian Fiskerstrand <k_f@g.o> wrote: |
| 4 |
|
| 5 |
|
| 6 |
> (up to discussion ofc) |
| 7 |
> |
| 8 |
> Pros for eblits vs the above eclasses: |
| 9 |
> - Let eclass/, which is a toplevel directory, be a place for code |
| 10 |
> useful to several packages, not a random dump of whatever people want |
| 11 |
> to share between ebuilds of the same package (yes, I'm looking at |
| 12 |
> you toolchain or apache2.eclass :) ). |
| 13 |
> - Eblits being in package directory, repoman checks can be extended to |
| 14 |
> look there. |
| 15 |
> - Have a guarantee that eblits are specific to a single package and |
| 16 |
> cannot "leak" to others by mistake: It greatly simplifies changing |
| 17 |
> and updating them. |
| 18 |
> - Eblits can be versionned, just the same as ebuilds or eclasses, but |
| 19 |
> old versions have a life expectancy more of the order of an ebuild |
| 20 |
> than that of an eclass. "Newborn" eblits would be expected to be |
| 21 |
> much more frequent than eclasses but less frequent than ebuilds. |
| 22 |
> - Eblits being per-package they'd obey to package rules, not eclass |
| 23 |
> rules wrt additions, removals, api-preservation, etc. |
| 24 |
|
| 25 |
This would be a policy question more than a technical one; we could |
| 26 |
decide e.g on a package-namespace[a] for eclasses following similar |
| 27 |
laxer rules[b]. |
| 28 |
|
| 29 |
> |
| 30 |
> Cons: |
| 31 |
> - Need a new, somewhat redundant, mechanism. |
| 32 |
> - Can be done without new EAPI but is then restricted to src_* phases. |
| 33 |
> - Very few consumers at the moment (though, considering the current |
| 34 |
> crusade that's not really a relevant metric to me). |
| 35 |
> |
| 36 |
|
| 37 |
Endnotes: |
| 38 |
[a] without changing any PMS (since review requirement is gentoo |
| 39 |
specific) it could be done by namespacing using hyphen or whatnot |
| 40 |
instead of separate directory. |
| 41 |
|
| 42 |
[b] to the extent more review isn't becoming the de-facto way forwards |
| 43 |
for all ebuilds anyways (we'd need better tooling) |
| 44 |
|
| 45 |
-- |
| 46 |
Kristian Fiskerstrand |
| 47 |
OpenPGP keyblock reachable at hkp://pool.sks-keyservers.net |
| 48 |
fpr:94CB AFDD 3034 5109 5618 35AA 0B7F 8B60 E3ED FAE3 |
Archives