1 |
On Sun, 2011-08-21 at 10:22 +1200, Kent Fredric wrote: |
2 |
> |
3 |
> On 21 August 2011 08:10, Matt Turner <mattst88@g.o> wrote: |
4 |
> |
5 |
> See https://bugs.gentoo.org/372513 |
6 |
> |
7 |
> ^ tldr version for everyone else. |
8 |
> |
9 |
> This is due to the || condition in virtual/fortran |
10 |
> |
11 |
> || ( sys-devel/gcc[fortran,openmp?] |
12 |
> sys-devel/gcc-apple[fortran,openmp?] dev-lang/ifc dev-lang/ekopath-bin |
13 |
> ) |
14 |
> |
15 |
> Where gcc[fortran] takes precedence over ifc. |
16 |
> |
17 |
> |
18 |
> |
19 |
> I wonder if there's some way we can manage this kind of |
20 |
> situation? |
21 |
> Perhaps portage could print alternative dependencies for |
22 |
> virtuals, |
23 |
> similar to the very helpful recent "The following keyword |
24 |
> changes are |
25 |
> necessary to proceed:" addition. |
26 |
> |
27 |
> |
28 |
> This usecases specifics aside, I'd welcome some sort of way to show |
29 |
> that an "or" condition is occurring somewhere in the tree, but it'd |
30 |
> have to be opt-in, instead of opt-out, as the potential for being very |
31 |
> noisy is great ( you'll get a lot of noise if you hit virtual/perl-* |
32 |
> for instance ). |
33 |
> |
34 |
> And likewise, I'd love to have "some way" to produce some sort of |
35 |
> graph for alternative merge trees that may work if you toggle some |
36 |
> variable, but the amount of complexity to do this I'd imagine is quite |
37 |
> large, and could easily be computationally expensive. |
38 |
> |
39 |
... |
40 |
|
41 |
> Alternatively, you could let the user dictate what type of |
42 |
> permutations to display/compute, ie: |
43 |
> |
44 |
> use-flag based permutations, keyword based permutations, mask-based |
45 |
> permutations, ||() conditional OR based permutations, |
46 |
> package-version/slot permutations etc. |
47 |
> |
48 |
> For "package version/slot" permutations, it would display every |
49 |
> variation on package/slot ( ie: slots/versions that are not the |
50 |
> "current" version ) that were installable, or installable with some |
51 |
> permutation ( if the depth of permutation is large enough ), so that a |
52 |
> user could see a path of installation they wanted and twist user masks |
53 |
> to make it happen. |
54 |
> |
55 |
> Of course, this is all looking like harder and harder stuff to do |
56 |
> ( programming is the gateway-drug for feature creep ) , but it would |
57 |
> still be something nice to have on a theoretical magic computer that |
58 |
> can do all computations in zero time. |
59 |
> |
60 |
> ¢¢ |
61 |
> -- |
62 |
> Kent |
63 |
> |
64 |
|
65 |
|
66 |
Well, for something now, you can emerge porthole and look at the |
67 |
Dependency view. It will display all options and show the recommended |
68 |
versions of the dep package (still needs some small tweaks to handle |
69 |
slots correctly). It is expandable (depth wise) until all deps are |
70 |
satisfied. The deps are also dbl-click able to pop up the dep in |
71 |
another window so you can view/change it's settings, merge, unmerge, |
72 |
etc.. |
73 |
|
74 |
And no it would not be suitable for portage, it is a guis based treeview |
75 |
and for human use only ;) |
76 |
|
77 |
But it can show you the alternatives for such cases. From there you can |
78 |
decide how you want to set things for the merges. |
79 |
-- |
80 |
Brian Dolbec <brian.dolbec@×××××.com> |