1 |
On Tue, 29 Apr 2003 19:44:41 +0000 |
2 |
Mikael Andersson <snikkt@×××××.com> wrote: |
3 |
|
4 |
> Gentoo is moving to slow this is why this kind of ideas emerge. It might be |
5 |
> moving fast, but obviously still too slow for many. I've got a few ebuilds |
6 |
> done which i haven't even bothered to try to get in since i've seen the queue |
7 |
> in bugzillla and I'm sure several others are in the same situation. |
8 |
|
9 |
Agreed .. |
10 |
|
11 |
> Communication between developers and users is always hard. Esp since many |
12 |
> gentoo users isn't strictly users. Some of the points i make might sound a |
13 |
> bit harsh - but try to smile while you read them - i think it will make |
14 |
> gentoo a better distribution if you do .) |
15 |
|
16 |
Personally .. I think a good "how to build ebuilds" would help alot if not a ebuild spacific |
17 |
mailling list. This way I can submit questions I have or ebuild for testing before submitting them to |
18 |
bugzilla ... |
19 |
|
20 |
I use existing ebuilds as a base for new ones since the skel in /usr/portage is really not that helpfull when you doin git for the first time. |
21 |
Maybe change skel can be a fully functional ebuils to do ummm .. nothing .. but have valid and correct ebuild structures and ways of doing things |
22 |
to help guide a new ebuilder in the right direction. |
23 |
|
24 |
I've had ebuilds that does wierd things that I don't understand .. and have asked for comments in bugzilla and on devel |
25 |
list but got no reply. |
26 |
|
27 |
this makes it hard to understand what I'm doing wrong or right .. |
28 |
|
29 |
> > Go back and read my posting about why many ebuilds don't get into the |
30 |
> > tree. |
31 |
> |
32 |
> I didn't see a lot of comments about the ebuild not being up to par in |
33 |
> bugzilla. Maybe if a developer seeing a build not beeing 'good-enough' could |
34 |
> add a small note about this so we - The Users(tm) - or in another word, the |
35 |
> user community [3] can fix whats wrong. We aim to please ... |
36 |
|
37 |
I think the suggestion of a ebuild mailling list for ebuild developers could be helpfull here as well. |
38 |
|
39 |
> No need to yell, instead add that information to [1]. In a way it's written in |
40 |
> [2] but as that is targeted for developers it's hard to know whether or not |
41 |
> I'm required to add that to my update, and if i should attach another file |
42 |
> with the entire changelog/changelog entry. And equally important, somebody |
43 |
> with the proper knowledge should get lintool online again. Apparently it's |
44 |
> broken ( see warnings in [2] ) and sure enough it reports errors for |
45 |
> baselayout and other core packages so it's probably at least somewhat broken. |
46 |
> > |
47 |
> > > I have come back to the new versions of the software... In a completely |
48 |
> > > new system, and the ebuild works fine. |
49 |
> > |
50 |
> > Is libcap even actively maintained still? If not and it is not widely |
51 |
> > used, then it is not likely to get into the tree at all. |
52 |
> |
53 |
> Maybe Larry the cow will stop using gentoo then ? After all he liked being in |
54 |
> control. If the users isn't in control gentoo will stop beeing bleeding edge |
55 |
> and maybe end up only bleeding. |
56 |
|
57 |
The time it takes to get feedback is discouraging ... I have helped with the ebuild for openexr and since then a new version has been released which I |
58 |
submitted and then an update to work with nv-cg-toolkit and the NVSDK and still no feedback ... |
59 |
|
60 |
I understand there are things happening to correct this and I'm waiting in antisipation for those changes as the only real |
61 |
contribution I can make to gentoo is bug reporting and ebuilds ... I'm doing lots of bug reporting ... but only one ebuild I had any interaction with has been |
62 |
incorporated into the tree. |
63 |
|
64 |
> > If you resolve all of the above issues, and the package is still widely |
65 |
> > used (I want some proof of this), then I'll take the ebuild and maintain |
66 |
> > it in portage myself. |
67 |
> |
68 |
> Well, this is a problem since not all packages wanting to be in portage will |
69 |
> be widely used. But i have hopes that these kind of problems will be resolved |
70 |
> with the restructuring of ebuild submission and maintainance in progress. |
71 |
|
72 |
If you meet a hungry man and give him a fish .. he eats for a day ... give him a fishing rod and he eats for a live time |
73 |
I knwo it's cliche'd .. but it's because it's true .. |
74 |
|
75 |
It will take alot less effort in the long run to help a "new" developer get it right and then monitor him to make sure he knows what |
76 |
he's doing then he can maintain he's own package. |
77 |
|
78 |
I do however thing a proper document on how to create correct ebuilds will go along way to helping ebuilders do things right. |
79 |
What is required etc etc. Sinply reading the emerge and ebuild documentation is not enought to understand the inner workings of the ebuild, |
80 |
and it has changed quiet a bit in the last few months. |
81 |
|
82 |
Once again I suggest a ebuild mailling list with the sole use of creating discussing and testing new ebuilds .. |
83 |
|
84 |
Henti |
85 |
|
86 |
-- |
87 |
gentoo-dev@g.o mailing list |