| 1 |
Marius Mauch wrote: |
| 2 |
|
| 3 |
> As a result of Cardoes earlier mail we talked a bit about possible |
| 4 |
> solutions in #gento-portage, and I suggested to let portage |
| 5 |
> automatically inject the deps based on SRC_URI pattern matching. |
| 6 |
> A mapping of extensions and their unpack deps would be kept in the tree |
| 7 |
> (e.g. mapping '.tar.bz2' to '( app-arch/tar app-arch/bzip2 )' |
| 8 |
> |
| 9 |
> Benefits: |
| 10 |
> - simplified depstrings for most packages (I assume >90% of the tree) |
| 11 |
> - easier maintenance if dependencies are changing, or additional |
| 12 |
> implementations become supported (this could also be achieved with |
| 13 |
> virtuals though) |
| 14 |
> - potentially more accurate dependencies, as some common errors would |
| 15 |
> be eliminated (e.g. proper treatment of use-conditionals, not adding |
| 16 |
> unpack deps to RDEPEND) |
| 17 |
> - long-term adds the possibility to remove bzip2, gzip and tar from |
| 18 |
> @system |
| 19 |
> |
| 20 |
> Potential problems: |
| 21 |
> - might cause trouble for some packages that use custom code for |
| 22 |
> unpacking, or due to circular deps, this could simply be solved with a |
| 23 |
> new RESTRICT value though. |
| 24 |
> - automagic deps could be confusing to devs/users |
| 25 |
> - not available for existing EAPIs (due to the mentioned problems) |
| 26 |
> - slightly slower dep calculations due to the extra processing |
| 27 |
> - possible match errors |
| 28 |
> |
| 29 |
> So, is this something ebuild maintainers would like in general, or does |
| 30 |
> such a feature cause you nightmares? |
| 31 |
|
| 32 |
Yes. I think that's something which should be done manually. |
| 33 |
|
| 34 |
|
| 35 |
|
| 36 |
-- |
| 37 |
gentoo-dev@l.g.o mailing list |
Archives