1 |
Christian Faulhammer wrote: |
2 |
> Doug Goldstein <cardoe@g.o>: |
3 |
> |
4 |
> |
5 |
>> Jonathan Adamczewski wrote: |
6 |
>> |
7 |
>>> Doug Goldstein wrote: |
8 |
>>> |
9 |
>>>> That's what this commits review list feels like. |
10 |
>>>> |
11 |
>>> Nearly every suggestion (from Donnie and others) has been over some |
12 |
>>> issue that relates directly to either correctness or |
13 |
>>> maintainability. It doesn't matter if you can "rattle off |
14 |
>>> capabilities to a millimeter" - if they're not documented somewhere |
15 |
>>> (like, say, in the comments of the ebuild) then the maintainer that |
16 |
>>> comes after you gets to go and break it all over again. |
17 |
>>> |
18 |
>> Correctness? Fine. Go ahead. Stick $(use_enable xvmc) into the ebuild. |
19 |
>> Do it. I dare you. Then try to compile. |
20 |
>> Guess what? When it blows up... that's called INcorrect. The opposite |
21 |
>> of the right thing. |
22 |
>> |
23 |
> |
24 |
> You were kindly asked if is not possible to use, so why do you feel |
25 |
> attacked? Do a comment on it and everybody would be fine, even the |
26 |
> people that would have to maintain it some time in the future. If you |
27 |
> don't like the review process, just ignore it. |
28 |
> Reviews are not a way to show what kind of idiot the committer is, but |
29 |
> to improve the overall quality of the tree. Nothing more, nothing less. |
30 |
> |
31 |
No. You clearly don't understand where I'm coming from. I think the |
32 |
commits review is pointless and a waste of resources that could be |
33 |
better used doing other things. Since commits review is a cyclic process |
34 |
you will never achieve a perfect state that all developers commit |
35 |
perfect ebuilds to the tree since new devs come and go. And since we |
36 |
don't document any of this stuff properly in the devmanual, incoming |
37 |
devs have to constantly relearn the same lessons that previous incoming |
38 |
devs learned through the review process. Effective workers work in 4 |
39 |
stages, we're effectively with this approach remaining in stage 1 and |
40 |
never progressing and admitting we will never progress. |
41 |
|
42 |
> |
43 |
> |
44 |
>> The maintainer who comes after me would be someone with a experience |
45 |
>> with the package. Some bumkin isn't going to come to maintain package |
46 |
>> XYZ unless they know or use the package, and guess what? That means |
47 |
>> experience. |
48 |
>> |
49 |
> |
50 |
> Yes, and the same goes for GNU Emacs, I needed some time to figure out |
51 |
> what all those things did and I broke it several times because I tried |
52 |
> to be clever. Now we documented it and I think everyone coming after us |
53 |
> will have a less hard time to understand it. Better document it, you |
54 |
> never know what happens. |
55 |
> |
56 |
> V-Li |
57 |
> |
58 |
> |
59 |
Read the ChangeLog. It's there for a reason. It provides valuable |
60 |
knowledge and information about the package. I would expect any |
61 |
developer worth their salt to at least brush up on the ChangeLog for any |
62 |
package they are taking over. |
63 |
|
64 |
-- |
65 |
gentoo-dev@g.o mailing list |