1 |
On 1.9.2011 14.31, Michał Górny wrote: |
2 |
> On Thu, 01 Sep 2011 14:02:11 +0300 |
3 |
> Petteri Räty <betelgeuse@g.o> wrote: |
4 |
> |
5 |
>> On 1.9.2011 13.51, Michał Górny wrote: |
6 |
>>> On Thu, 01 Sep 2011 13:44:47 +0300 |
7 |
>>> Petteri Räty <betelgeuse@g.o> wrote: |
8 |
>>> |
9 |
>>>> On 1.9.2011 12.03, Michał Górny wrote: |
10 |
>>>>> Hello, |
11 |
>>>>> |
12 |
>>>>> A quick idea. Right now eclasses sometimes do API changes and |
13 |
>>>>> start yelling at users merging ebuilds using outdates APIs. This |
14 |
>>>>> often means users start filling bugs about outdated ebuilds |
15 |
>>>>> requiring maintainers either to ignore that or start updating old |
16 |
>>>>> ebuilds retroactively. |
17 |
>>>>> |
18 |
>>>>> Maybe we should add some kind of devqawarn() function to |
19 |
>>>>> eutils.eclass, which would trigger special QA warnings only when |
20 |
>>>>> ebuild is merged by a developer? This could be triggered e.g. by |
21 |
>>>>> some kind of voluntary make.conf setting. |
22 |
>>>>> |
23 |
>>>> |
24 |
>>>> What's wrong with eqawarn that's already provided by eutils? |
25 |
>>> |
26 |
>>> The first paragraph? |
27 |
>>> |
28 |
>> |
29 |
>> Have Portage defaults so that users only see if them if they read the |
30 |
>> merge logs and then developer profiles can set the settings to log |
31 |
>> them? |
32 |
> |
33 |
> 1) that's changing existing behavior, |
34 |
> 2) what with non-portage users? |
35 |
> |
36 |
|
37 |
1) eqawarn == devqawarn. I don't see a reason to come up with a new |
38 |
function just to avoid changing Portage configuration. |
39 |
|
40 |
2) How messages from each e* function is shown is left to the package |
41 |
manager to decide. |
42 |
|
43 |
One thing to note is that we should get eqawarn into the next EAPI. |
44 |
|
45 |
Regards, |
46 |
Petteri |