1 |
Ciaran McCreesh schrieb: |
2 |
> On Sat, 23 Apr 2011 16:47:37 +0200 |
3 |
> Chí-Thanh Christopher Nguyễn <chithanh@g.o> wrote: |
4 |
> |
5 |
>> What I propose solves the problems that old-style virtuals introduce |
6 |
>> in dependency resolution. |
7 |
>> |
8 |
> Not really, because it means we'd have to keep the nasty old code around |
9 |
> forever. If we just do away with the things entirely instead then we |
10 |
> can pretend they never existed (like we did for ? : dependencies). |
11 |
> |
12 |
|
13 |
We can mostly pretend that they never existed. PROVIDE will only be |
14 |
relevant during the installation phase. |
15 |
|
16 |
>> What other problems do they cause? |
17 |
>> |
18 |
> DEPEND=">=virtual/blah-2" |
19 |
> |
20 |
> DEPEND="virtual/blah[foo]" |
21 |
> |
22 |
> DEPEND="!virtual/that-i-provide" |
23 |
> |
24 |
> PROVIDE="not-a/virtual" |
25 |
> |
26 |
Would be ignored during dependency calculation and become |
27 |
not-a/virtual-<version> in package.provided later. |
28 |
> best_version virtual/blah |
29 |
> |
30 |
> The full VDB load required to figure out whether or not a virtual is |
31 |
> installed |
32 |
|
33 |
Problems I had indeed not taken into account are !virtual/foo |
34 |
dependencies and USE flags in package.provided (not allowed yet, bug |
35 |
142941). But the former are also not going to work any more if new-style |
36 |
virtuals are used instead of old-style, as you block only the virtuals |
37 |
and not their providers. |
38 |
|
39 |
If the USE flag problem is considered a show-stopper, then I have to |
40 |
take back my proposal until bug 142941 is fixed. |
41 |
|
42 |
|
43 |
Regards, |
44 |
Chi-Thanh Christopher Nguyen |