1 |
Can anyone point me to any docs on, or provide any insight into, the |
2 |
algorithm emerge uses to resolve DEPEND requests? I can't make heads or |
3 |
tails of the results I'm getting from "emerge system". |
4 |
|
5 |
|
6 |
I'm maintaining a set of ebuilds locally for a number of internal |
7 |
packages. I am trying to control them as a group using a single ebuild |
8 |
"wrapper" that depends on them; version 1.0 of the wrapper would have: |
9 |
|
10 |
DEPEND=" ~local/xxx-1.1 ~local/yyy-1.4 ~local/zzz-2.0 " |
11 |
|
12 |
then version 2.0 of the wrapper would have different versions. Then I |
13 |
put the wrapper into my packages file and when I run "emerge system", I |
14 |
should get what I want. |
15 |
|
16 |
Internally these packages have their own DEPEND lines, and they may well |
17 |
be less restrictive; local/xxx-1.1 might say it depends on "local/zzz" |
18 |
(any version) for example. |
19 |
|
20 |
|
21 |
But, it doesn't work the way I'd expect. I assumed emerge would examine |
22 |
all the DEPEND requests and come up with a list of packages which met |
23 |
all the criteria (or give an error if there wasn't one). |
24 |
|
25 |
Instead what seems to happen is that emerge sees the DEPEND with no |
26 |
version in local/xxx-1.1 and pulls in the latest local/zzz. Then it |
27 |
proceeds along and sometime later it discovers that it really wanted an |
28 |
earlier version so it installs that instead. |
29 |
|
30 |
Not only that, but I tried to arrange my packages in the wrapper DEPEND |
31 |
so that the ones that didn't depend on anything were first, etc. so |
32 |
that, if emerge did them in order, I'd get the right versions in the |
33 |
end. I didn't expect to have to do this but I tried it anyway; no |
34 |
help. Emerge merges them in some seemingly random order anyway. |
35 |
|
36 |
|
37 |
I'm using Portage 2.0.49. |
38 |
|
39 |
Can someone give me some tips/help/pointers/debugging advice/anything? |
40 |
|
41 |
I'm pretty stuck here :-(. |
42 |
|
43 |
-- |
44 |
------------------------------------------------------------------------------- |
45 |
Paul D. Smith <psmith@××××××××××××××.com> HASMAT: HA Software Mthds & Tools |
46 |
"Please remain calm...I may be mad, but I am a professional." --Mad Scientist |
47 |
------------------------------------------------------------------------------- |
48 |
These are my opinions---Nortel Networks takes no responsibility for them. |
49 |
|
50 |
-- |
51 |
gentoo-dev@g.o mailing list |