| 1 |
-----BEGIN PGP SIGNED MESSAGE----- |
| 2 |
Hash: SHA1 |
| 3 |
|
| 4 |
Marius Mauch wrote: |
| 5 |
> Just ran across the following thread in the forums yesterday: |
| 6 |
> http://forums.gentoo.org/viewtopic-t-626528.html |
| 7 |
> |
| 8 |
> Which raises an interesting point regarding merge order of runtime |
| 9 |
> deps. IIRC we currently assume that it's ok to merge runtime deps after |
| 10 |
> the depending package to resolve dep cycles for example, which is |
| 11 |
> generally ok, except if a runtime dep is used in pkg_*. For |
| 12 |
> ebuild-installs that can be worked around easily by using DEPEND (where |
| 13 |
> order is strictly respected), but for binary packages that obviously |
| 14 |
> doesn't work. |
| 15 |
> This problem probably hasn't been recognized earlier as |
| 16 |
> it requires several conditions to apply simultaneously (binary merge, |
| 17 |
> circular rdeps, rdeps used in pkg_*, rdeps not installed |
| 18 |
> previously) |
| 19 |
> |
| 20 |
> Assuming I haven't missed anything, I see threee options to deal with |
| 21 |
> that problem: |
| 22 |
> a) ignore it, as it only affects a small minority |
| 23 |
> b) respect merge order for RDEPEND - will cause more unsolvable |
| 24 |
> depgraphs, though telling people to use PDEPEND more often might reduce |
| 25 |
> that problem |
| 26 |
|
| 27 |
The resolver currently tries to merge both RDEPEND and PDEPEND |
| 28 |
before whenever possible. There is an optimization in 2.1.4_rc that |
| 29 |
improves merge order in some circular RDEPEND cases, see the |
| 30 |
cmp_circular_bias() function in depgraph.altlist(). There was |
| 31 |
another related optimization for bug #189966 that's already in 2.1.3.19. |
| 32 |
|
| 33 |
I would encourage people to use PDEPEND whenever appropriate. Since |
| 34 |
the fix for bug 176765 (2.1.2.6) it behaves very similar to |
| 35 |
RDEPEND, so it should be usable in more cases. |
| 36 |
|
| 37 |
> c) add a new deptype for merge dependencies - looks like overkill to me |
| 38 |
|
| 39 |
Actually, a new dep type seems pretty reasonable to me. We can |
| 40 |
consider it part of bug 174552. |
| 41 |
|
| 42 |
Zac |
| 43 |
|
| 44 |
> Any other other ideas, comments, preferences? |
| 45 |
> |
| 46 |
> Marius |
| 47 |
|
| 48 |
-----BEGIN PGP SIGNATURE----- |
| 49 |
Version: GnuPG v2.0.7 (GNU/Linux) |
| 50 |
|
| 51 |
iD8DBQFHVuxy/ejvha5XGaMRAl81AJ4w9AxdA1s3TumsHd6QW18NXl6YXACgshlg |
| 52 |
kTRZh6u5neywMTH5cPaGcSM= |
| 53 |
=DFSH |
| 54 |
-----END PGP SIGNATURE----- |
| 55 |
-- |
| 56 |
gentoo-portage-dev@g.o mailing list |
Archives