1 |
On 2002.11.09 14:38 George Shapovalov wrote: |
2 |
> Hey Evan, Javier and everybody. |
3 |
> |
4 |
> So this issue came up again, yet another time (TM) :). |
5 |
> Please take a look at #1523. That describes some thoughts and certain |
6 |
> strategy |
7 |
> that is designed to help exactly this. |
8 |
|
9 |
Interesting. Your thoughts on offloading "non-critical" tasks to users |
10 |
probably is the solution much more than changing the infrastructure. |
11 |
|
12 |
My beef is the length of time things sit in bugzilla waiting for |
13 |
inclusion. Are there meant to be numerous comments on the bug talking |
14 |
of success before things will get into the tree? |
15 |
|
16 |
> the now-ultimate solution will involve streamlined ebuild |
17 |
> sbmission/processing that will allow fast turnover and wide-scale |
18 |
> testing |
19 |
> that you are proposing here. Comments are wellcome (specifically the |
20 |
|
21 |
How about this for a suggestion: |
22 |
|
23 |
I know that nobody wants to duplicate effort etc, but would it be |
24 |
possible to break the portage tree up (one tree, many dirs)? |
25 |
|
26 |
Maybe I am thinking too much about the debian scenario, but what do |
27 |
people think of having something like: |
28 |
|
29 |
/usr/portage/stable |
30 |
/usr/portage/testing |
31 |
/usr/portage/unstable |
32 |
|
33 |
This would be a single tree, but a new option (which would remove |
34 |
keywords, I guess - sorry ;) in make.conf (or wherever) would be |
35 |
"DEFAULT_TREE=". |
36 |
|
37 |
So one could track the stable, testing or unstable tree. To select a |
38 |
package outside your default tree would be something like: |
39 |
|
40 |
# emerge --unstable bogofilter |
41 |
|
42 |
There would be duplication of ebuilds, but they are only small |
43 |
anyways. Different versions would exist in different places. If there |
44 |
was only a stable version, and no testing versions where available, |
45 |
then something like |
46 |
|
47 |
"Package not found in unstable tree" would come up. |
48 |
|
49 |
Users could submit ebuilds (like my judy and bogofilter - and the |
50 |
folding@home one I am thinking of doing) that could be checked in by a |
51 |
core developer after they have verified it isn't majorly screwed. In |
52 |
otherwords, it hits the tree if it builds and installs (even if rough - |
53 |
price you pay for --unstable). |
54 |
|
55 |
Then it could sit in unstable where people could test it and put |
56 |
suggestions against it. We could use existing infrastructure, things |
57 |
would get into the tree so much faster, and masking and keywords |
58 |
wouldn't be required. By all means, append comments to buzilla bugs to |
59 |
indicate how good the package is. |
60 |
|
61 |
The idea isn't to change the way things are done, just provide a way to |
62 |
get stuff into the tree. |
63 |
|
64 |
> In the mean time it is possible to announce new ebuilds here, how |
65 |
> this is |
66 |
> done now. |
67 |
|
68 |
Possible to auto mate the announcement of ebuilds from new bugs in |
69 |
buzilla under "ebuild" category? Subscribe to |
70 |
new-ebuild-announce@g.o? |
71 |
|
72 |
> Hereby I would like to ask anybody willing to do some testing (and not |
73 |
> afraid |
74 |
> of bad effects of course ;)), to try out new ebuilds which have "~" |
75 |
> for their |
76 |
|
77 |
So, enable: #ACCEPT_KEYWORDS="~arch"? |
78 |
|
79 |
Perhaps the developers could publish a checklist of things that users |
80 |
could do? |
81 |
|
82 |
So in bugzilla, they could comment on each thing and say "works" or |
83 |
"fails". Things like successful install, removal etc to test the |
84 |
ebuild functions springs to mind. If rc scripts come with it, checks |
85 |
for stopping and starting successfully could also work. |
86 |
|
87 |
Also a suggestion to check for upstream bugs for anything release |
88 |
critical. |
89 |
|
90 |
Developers, teach us ;) |
91 |
|
92 |
> ARCH in the KEYWORDS and report results. The best way to report at |
93 |
> present is |
94 |
> to find the related bug (just a quick search on bugzilla) and post a |
95 |
> message |
96 |
> there (do not be afraid if the bug is closed - core developer will be |
97 |
> glad to |
98 |
> hear about success and will remove "~" onces he accumulates anough |
99 |
> feedback). |
100 |
|
101 |
If that feedback could be documented, and the process required defined, |
102 |
that would be great. |
103 |
|
104 |
It would also be great if a statement about "this is what keeps ebuilds |
105 |
out of the tree" would also be good. What is needed to satisfy the |
106 |
core team? |
107 |
|
108 |
Thanks. |
109 |
|
110 |
Evan. |
111 |
|
112 |
-- |
113 |
gentoo-dev@g.o mailing list |