1 |
-----BEGIN PGP SIGNED MESSAGE----- |
2 |
Hash: SHA256 |
3 |
|
4 |
On 11/18/2015 01:54 AM, Raymond Jennings wrote: |
5 |
> On Wed, Nov 18, 2015 at 1:25 AM, Alexander Berntsen |
6 |
> <bernalex@g.o <mailto:bernalex@g.o>> wrote: |
7 |
> |
8 |
> On 18/11/15 08:25, Ulrich Mueller wrote: |
9 |
>> - If you mix stable and unstable then you are by definition an |
10 |
>> advanced user, who will be able to cope with the situation. :) |
11 |
> This attitude is shitty, and I am willing to wager a really big |
12 |
> bunch of users fall into this category. |
13 |
> |
14 |
> |
15 |
>> I second the motion. |
16 |
> |
17 |
>> I can vouch for myself as such a user who often uses unstable |
18 |
>> packages to get new features or dodge bugs in the stable |
19 |
>> versions. |
20 |
> |
21 |
>> As a general default, though, I stick with stable packages. |
22 |
> |
23 |
>> Taking an occasional unstable package in an otherwise stable |
24 |
>> system (and obeying any revbump directives provoked by |
25 |
>> dependencies) is a legal operation that is actually supported by |
26 |
>> commit rules that prohibit stable versions from depending on |
27 |
>> unstable versions of dependencies. |
28 |
> |
29 |
>> Why would that policy exist if mixing stable and unstable were |
30 |
>> unsupported? |
31 |
> |
32 |
|
33 |
Mixing keywords is generally unsupported because of the sheer amount |
34 |
of testing and possible combinations there are. In many cases, you can |
35 |
get away with it and it's easy enough to debug, but if you end up in |
36 |
package.use hell and/or have a bunch of blockers, you're kinda on your |
37 |
own. |
38 |
|
39 |
The policy exists to prevent developers from causing users to fall |
40 |
into the trap of mixing keywords. Stable packages depending on testing |
41 |
packages means they'll have to add entries to package.accept_keywords, |
42 |
which can greatly complicate maintenance of a Gentoo system. Someone |
43 |
willing to do that work *is* advanced, even if it's only a few |
44 |
packages. "Advanced" doesn't have to mean "can build an LFS system in |
45 |
an evening". :) |
46 |
|
47 |
In the case of this recent change, worst case scenario you'll need to |
48 |
add portage to your p.accept_keywords until it's stable. If you want |
49 |
to simplify things, choose arch or ~arch without mixing. |
50 |
|
51 |
In the Gentoo spirit, though, users are free to do with their systems |
52 |
whatever they want; mixing keywords is one of the cases where you get |
53 |
to keep the pieces, however. |
54 |
|
55 |
> I don't think EAPI 6 is *that* shiny, that we need to start using |
56 |
> it prior to stable Portage supporting it. It's a potential mess for |
57 |
> a huge portion of our users. |
58 |
> |
59 |
> |
60 |
> That said, I'd like to extend my thanks to Micha? for working on |
61 |
> this. |
62 |
> |
63 |
> |
64 |
|
65 |
- -- |
66 |
Daniel Campbell - Gentoo Developer |
67 |
OpenPGP Key: 0x1EA055D6 @ hkp://keys.gnupg.net |
68 |
fpr: AE03 9064 AE00 053C 270C 1DE4 6F7A 9091 1EA0 55D6 |
69 |
-----BEGIN PGP SIGNATURE----- |
70 |
Version: GnuPG v2 |
71 |
|
72 |
iQIcBAEBCAAGBQJWTYSSAAoJEAEkDpRQOeFw0w0QANEwsIEHWWmZ0jhHJTGHZY1s |
73 |
oYogl3uNYQypSchyUait6VBMfTCvQ/mOCVgZMy8iIegKRbpY7tWqFG5GA2nXW0br |
74 |
FRXhsrC5hg5pFDKRj3cpI2DURXWrFomRs+fKrEbpOvy5EeTEcki8o1+tRb7D2FWj |
75 |
xGT8hkIbs3i2d1wSBzr2HPJEm7DS0259LrrmcT4EPKp6fuRZzAT3YScSN3f7wFBg |
76 |
OrFZwtlW4MWQSmHhq0z0MqnG50y6j6xIlECwUCo/zGjVn/7ergd7BXldXb18ix11 |
77 |
BJgqWkslIf4Ys6jSgzcecTTDF6eimi6Owc05PhOHqMdH/iU2uPm7RekmJGRaUWel |
78 |
jwYutEpnzMR2dyxWlPVWL+O34UFTMKpw/BFwULg6zSBJpbYfwfFvPoWlLSTv/w1m |
79 |
oikrjXiQArP2Hr7pL/Y+E2z/acFQeKg1oNIFOWac6UEyh35rfCZErdw+zLwGjU7A |
80 |
RYtXWWzbfHeBu3lNIYWKIpvov0BOySXJ6R/rinxvyfO0ojr1WveOajRgGGgjl7Sq |
81 |
BfLG0SMhfY5qOXqw2NRARB8UFst96k/aPqPLXYdoOeIr97hx8tPurKFy1DXyHaxM |
82 |
Jyox5844Fv9hbPaAIxc+S8EppslnExp8G/bM7OuLVdy1/H0iKOt6npI7OcOcMKt/ |
83 |
6rQ4FVxXoxGzayM4yUYq |
84 |
=83f3 |
85 |
-----END PGP SIGNATURE----- |