1 |
On Sun, 20 Jan 2013 23:33:39 +0100 |
2 |
Michał Górny <mgorny@g.o> wrote: |
3 |
|
4 |
[...] |
5 |
> > Do you plan to keep precise depends for packages? |
6 |
> > like glibc[abi_x32]/gcc[abi_x32] for all libraries requesting x32. |
7 |
> |
8 |
> Yes. ${MULTILIB_USEDEP} is for that (it currently evaluates |
9 |
> to 'multilib?'). |
10 |
|
11 |
In that very precise case (gcc/glibc) I'd say no: it's probably saner |
12 |
to leave the toolchain as it is, ie, able to build all supported abis. |
13 |
It probably means an end to implicit system deps for the rest of the |
14 |
system set though. |
15 |
|
16 |
[...] |
17 |
> > like on ABI=amd64 media-libs/glu[ABI=x32] could not be used by |
18 |
> > any of ABI=amd64 users. |
19 |
> > |
20 |
> > In order to track such depends precisely you would need to add |
21 |
> > ABI flags to each revdep recursively. It's quite invasive. Is it |
22 |
> > worth the effort? |
23 |
> |
24 |
> A good point. I'd say that the default impl should be built then. |
25 |
> But... how about making it a USE flag with use.force logic? That way, |
26 |
> it would be explicitly visible, and if someone really wanted to |
27 |
> disable it, he would be able to do it on his own responsibility. |
28 |
|
29 |
+1 |
30 |
|
31 |
[...] |
32 |
> > Looks like insane amount of metadata growth for each |
33 |
> > plagued package. |
34 |
> |
35 |
> I don't think metadata is really important here. I believe that |
36 |
> the amount of additional metadata introduced by the packages affected |
37 |
> by multilib is not really one worth worrying. |
38 |
> |
39 |
|
40 |
+1 too |
41 |
OTOH, if you don't have this metadata you cannot really distinguish |
42 |
between packages needing multilib and those that do not care (I |
43 |
wouldn't want to run texlive-latexextra src_* phases twice or three |
44 |
times because I want multilib). |
45 |
|
46 |
Alexis. |