1 |
"Tod M. Neidt" <tod@g.o> writes: |
2 |
|
3 |
> Hi! |
4 |
|
5 |
arp. :) |
6 |
|
7 |
> I'll take a stab at this and offer some suggestions. Please note |
8 |
> that these comments should not be considered "policy" and it should |
9 |
> be a given that any drobbins pronouncements on this issue |
10 |
> automatically trump mine. |
11 |
|
12 |
okidoki. thanks. |
13 |
|
14 |
> On Thu, 2002-04-11 at 15:30, Terje Kvernes wrote: |
15 |
> |
16 |
> [ ... ] |
17 |
> |
18 |
> > first off, if I see an ebuild that is out of date, do I contact |
19 |
> > the author of the out-of-date ebuild or do I submit a new one? so |
20 |
> > far I've thought "if the old one is _really_ out of date, I'll |
21 |
> > submit a new ebuild, if it's a week or two since the new version |
22 |
> > came out, I'll take it with the author." the real question is |
23 |
> > probably how "possessive" authors are / are supposed to be with |
24 |
> > regards to their ebuilds. is there any set policy on this? |
25 |
> |
26 |
> I would suggest submitting any changes to bugzilla and notifying the |
27 |
> "author" or "maintainer" by cc'ing them on the bug submission if |
28 |
> they don't happen to be the automatic assignee. |
29 |
|
30 |
good point. |
31 |
|
32 |
> Note that you should probably check the Changelog for that |
33 |
> particular package to see who has been actively maintaining the |
34 |
> ebuild, as this developer is not necessarily the same as the one |
35 |
> listed in the ebuild header. |
36 |
|
37 |
that much I figured on my own, but it is none the less good advice. |
38 |
besides, this happens to be stored in the list archives, so |
39 |
additional information is always a good thing[tm]. |
40 |
|
41 |
> If the ebuild update just requires a copy of the ebuild to the new |
42 |
> revision number, it is not necessary to attach the ebuild. A |
43 |
> comment stating that "version 1.2 of foo has been released,a simple |
44 |
> copy of foo-1.1-r3.ebuild to foo-1.2.ebuild works" or something to |
45 |
> that effect is sufficient. |
46 |
|
47 |
oki. |
48 |
|
49 |
> If the ebuild update requires changes to the ebuild, I find it |
50 |
> convenient when people attach a diff between the original and |
51 |
> updated ebuild instead of the entire updated ebuild as this makes it |
52 |
> very easy to see the changes that are proposed. |
53 |
|
54 |
good point again. I wasn't all that comfortable with creating diffs |
55 |
until I wrote the xgammon ebuild yesterday. then I didn't have much |
56 |
choice. :) |
57 |
|
58 |
> > second, and a bit more tricky. I thought I'd make an xgammon |
59 |
> > ebuild. all fine and dandy, but there are a few patches (from |
60 |
> > RedHat) to xgammon which are nice to have (they fix minor bugs -- |
61 |
> > void main -> int main and a few other things). what to do? if I |
62 |
> > include the patches, how do I link to them? or do I just put them |
63 |
> > in the "files"-directory with the ebuild? |
64 |
> |
65 |
> For patch files (especially large ones) include the URL to the patch |
66 |
> in the SRC_URI string so that it will be downloaded with the source |
67 |
> tarball and stored in ${DISTDIR}, i.e. /usr/portage/distfiles by |
68 |
> default. See the dev-lang/python ebuild for a good example (In fact |
69 |
> the python ebuild is a good example for a variety ebuild techniques |
70 |
> for uncommon situations) |
71 |
|
72 |
hm. the patches I need total about 1K, and I _could_ host the |
73 |
patches via http. |
74 |
|
75 |
> If the patch files are small (working definition of small not much |
76 |
> larger than the ebuild itself), placing them in the files directory |
77 |
> is ok, but the former method is prefereable. |
78 |
|
79 |
okay, I'll try getting them to work via http. <pause for half an |
80 |
hour>. okay, done. :) now just to check that things work and |
81 |
submit. should I attach the patches as well? |
82 |
|
83 |
> Contributions are always welcome and appreciated. |
84 |
|
85 |
considering how much the Gentoo community (well, the linux community |
86 |
in general actually) has given me I feel bad for not contributing |
87 |
more than I do. :/ |
88 |
|
89 |
> Hope that helps, |
90 |
|
91 |
always, thank you. |
92 |
|
93 |
-- |
94 |
Terje |