1 |
> On 30/09/12 06:15 PM, Brian Harring wrote: |
2 |
> |
3 |
>>>>> yngwin has a point that I've not seen addressed. |
4 |
>>>>> |
5 |
>>>>> What /is/ wrong with the whole CDEPEND intermediate var idea? |
6 |
>>>>> |
7 |
>>>> |
8 |
>>>> The problem appears as we introduce more DEPEND variables |
9 |
>>>> (which is what prompted the proposal, IIRC). If we have |
10 |
>>>> ADEPEND, BDEPEND, CDEPEND, and DDEPEND, and there's only some |
11 |
>>>> (i.e. not total) sharing going on then the COMMON_DEPEND |
12 |
>>>> pattern starts to fall apart. You potentially need, |
13 |
>>>> |
14 |
>>>> AB_DEPEND AC_DEPEND AD_DEPEND BC_DEPEND BD_DEPEND CD_DEPEND |
15 |
>>>> ABC_DEPEND ABD_DEPEND ACD_DEPEND BCD_DEPEND ABCD_DEPEND |
16 |
>>>> (COMMON_DEPEND) |
17 |
>>>> |
18 |
>>>> This obviously gets worse as more DEPEND vars are introduced. |
19 |
>>>> |
20 |
>>> |
21 |
>>> Well not really, no -- the additional *DEPENDs that are being |
22 |
>>> proposed (or at least mentioned) for new EAPI will either remove |
23 |
>>> atoms from COMMON_DEPEND/DEPEND/RDEPEND or will be used so |
24 |
>>> tersely that a COMMON_DEPEND or other intermediate variable won't |
25 |
>>> really be necessary for them. |
26 |
|
27 |
Another thing I wanted to point out is that those "potential" extra |
28 |
variables are not needed in practice. We already have 98% of the tree |
29 |
(if I got the previously mentioned stats right) that does fine with |
30 |
just one or two ({R,}DEPEND). The majority of that other 2% needs just |
31 |
one more variable. There may be corner cases where more vars would be |
32 |
needed, but those will never be more than a few ebuilds. |
33 |
|
34 |
It's just not worth it to completely change the way we do things (or |
35 |
use two systems in parallel) just for a few ebuilds that would |
36 |
significantly benefit. |
37 |
|
38 |
If we were a new distro and designing our ebuild format from scratch, |
39 |
then yes, I would say your proposal has merit. But we aren't. We have |
40 |
hundreds of people and tens of thousands of ebuilds using *DEPEND just |
41 |
fine. There are no big problems, only corner-cases. We're not talking |
42 |
about incremental improvements either (such as was the case e.g. with |
43 |
use deps). |
44 |
|
45 |
Let's just keep things simple, and refrain from "fixing" what isn't broken. |
46 |
-- |
47 |
Cheers, |
48 |
|
49 |
Ben | yngwin |
50 |
Gentoo developer |
51 |
Gentoo Qt project lead, Gentoo Wiki admin |