1 |
On 01/25/2014 01:09 PM, Chris Reffett wrote: |
2 |
> On 01/25/2014 05:12 AM, Markos Chandras wrote: |
3 |
>> On 01/23/2014 04:48 PM, Michał Górny wrote: |
4 |
>>> Dnia 2014-01-23, o godz. 11:36:06 Chris Reffett |
5 |
>>> <creffett@g.o> napisał(a): |
6 |
>>> |
7 |
>>>> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 |
8 |
>>>> |
9 |
>>>> On 01/23/2014 11:28 AM, Michał Górny wrote: |
10 |
>>>>> Dnia 2014-01-23, o godz. 11:24:41 Chris Reffett |
11 |
>>>>> <creffett@g.o> napisał(a): |
12 |
>>>>> |
13 |
>>>>>> After some discussion on good ways to communicate optional |
14 |
>>>>>> dependencies to users, I was shown the optfeature() |
15 |
>>>>>> function in net-misc/netctl. Gentoo contributor Andrew |
16 |
>>>>>> Hamilton and I came up with a cleaned up and expanded |
17 |
>>>>>> version of it, and I would like to add it to eutils.eclass |
18 |
>>>>>> to provide a standard way of notifying users of optional |
19 |
>>>>>> dependencies. The patch to eutils.eclass is attached. |
20 |
>>>>> |
21 |
>>>>> This was discussed already: |
22 |
>>>>> |
23 |
>>>>> http://thread.gmane.org/gmane.linux.gentoo.devel/72162 |
24 |
>>>>> |
25 |
>>>> First of all, this is a short patch for a function, not a full |
26 |
>>>> eclass. |
27 |
>>> |
28 |
>>> Ah, sorry, this changes *a lot*. Let's start the bikeshed again |
29 |
>>> then, whatever. |
30 |
>>> |
31 |
>> I haven't looked at the implementation, but I wonder if we need a |
32 |
>> function for such trivial stuff. Most maintainers deal with this |
33 |
>> problem using pkg_postinst() einfo/elog messages. Why do we need a |
34 |
>> dedicated function for that? Just for consistency reasons...? |
35 |
> |
36 |
> Consistency, and because it removes the need for a bunch of "if |
37 |
> has_version" lines, instead only displaying if you don't satisfy the |
38 |
> deps (and supports both "and" and "or" groupings for packages |
39 |
> satisfying the dep). This also stems from a complaint I've seen a lot |
40 |
> about how optional dep messages should only display if the requisite |
41 |
> package isn't installed, this makes that job a little simpler. But |
42 |
> mostly consistency, this gives us one nice function that we can |
43 |
> standardize on. |
44 |
> |
45 |
> Chris Reffett |
46 |
> |
47 |
|
48 |
I am fine with that especially given it's an opt-in feature. |
49 |
|
50 |
-- |
51 |
Regards, |
52 |
Markos Chandras |