1 |
On Sun, Sep 20, 2015 at 7:52 AM, Neil Bothwick <neil@××××××××××.uk> wrote: |
2 |
> On Sat, 19 Sep 2015 20:37:53 -0400, Philip Webb wrote: |
3 |
> |
4 |
>> My impression is that using Portage has become more complicated |
5 |
>> & its warning/error messages have not been given the necessary |
6 |
>> attention. Complaints or pleas for help like the OP's here are quite |
7 |
>> frequent & not all of them come from novices who don't understand what |
8 |
>> Gentoo does. |
9 |
>> |
10 |
>> Portage sb able to report eg |
11 |
>> |
12 |
>> "Pkg1 & Pkg2 (of Version1 & Version2) can't be installed together ; |
13 |
>> Pkg1 is needed for Pkg3, which you already have installed ; |
14 |
>> Pkg2 is needed for Pkg4, which you are trying to install. |
15 |
>> Please review your needs : you may need to remove a package |
16 |
>> temporarily in order for Portage to proceed, then restore a different |
17 |
>> version of it". |
18 |
> |
19 |
> Maybe it should, but if there is no one willing or able to take on this |
20 |
> task, it won't. |
21 |
|
22 |
So, kicking the overworked portage team with stuff like "Gentoo has a |
23 |
lousy package manager" is not helpful and certainly violates the CoC. |
24 |
I don't see that here. |
25 |
|
26 |
On the other hand, that doesn't mean that we all need to line up and |
27 |
drink the kool aide and say that the behavior pointed out in the |
28 |
original message is desired behavior. |
29 |
|
30 |
We can acknowledge that bugs exist without lining up with signs |
31 |
demanding their immediate fix. The portage team does great work, but |
32 |
the fact that package runtime dependencies can vary so much compared |
33 |
to a binary distro greatly complicates the dependency-resolution |
34 |
problem. So do some of our package-maintenance practices, and those |
35 |
are being looked at right now. |
36 |
|
37 |
Something outsiders probably could contribute might be something like |
38 |
a guide to portage troubleshooting on the wiki that lists some common |
39 |
scenarios and their workarounds, or possibly working with the portage |
40 |
team to get short references to such a guide added to the portage |
41 |
output. So, portage might suggest re-running it with --backtrack=# or |
42 |
whatever if it outputs the sort of errors that backtracking is likely |
43 |
to fix, and so on. Just doing that alone would probably triage a |
44 |
large number of issues that confuse users which makes them somewhat |
45 |
happier and cuts down on list traffic. |
46 |
|
47 |
-- |
48 |
Rich |