1 |
On Fri, Mar 13, 2009 at 8:52 AM, Marcelo Andrier |
2 |
<marcelo.andrier@×××××.com> wrote: |
3 |
> I failed to see why the "Cache sync" idea is too ambitious. It's not |
4 |
> only about space but also speed. It still takes forever to sync and |
5 |
> search packages, and really one does not need all the cruft from the |
6 |
> portage tree. Besides I wonder how many Gentoo users have their |
7 |
> portage directory on squashfs, and users with space limits usually |
8 |
> have other limits. I bet a cache sync would actually be much more |
9 |
> welcome than a database for file collisions. Students and mentors, do |
10 |
> consider! |
11 |
|
12 |
What are you using to search? esearch for instance takes about .7 |
13 |
seconds to search the entire tree for a package. I bet eix is as fast |
14 |
or faster. |
15 |
emerge -s is probably slow. It might be useful to pragmatic to just |
16 |
remove it and point users at tools that don't take forever. |
17 |
> |
18 |
> Also please don't remove the ideas on the wiki. Even if one dev |
19 |
> doesn't think it's a good idea does not mean someone else won't. The |
20 |
> fact it was the first idea a student picked meant it's not so bad, |
21 |
> even if it could be improved. |
22 |
|
23 |
It has nothing to do with being the first idea. If people add stupid |
24 |
ideas to the wiki you bet we will call them out as such when people |
25 |
want to work on them. Trust me when I say about half of the |
26 |
applications we will likely receive will be about an idea that is: |
27 |
|
28 |
1) so poorly specified we can't figure out what the actual project is |
29 |
2) already implemented somewhere else |
30 |
3) just technically infeasible |
31 |
4) Not worth the effort |
32 |
5) Not able to be completed in 3 months |
33 |
6) able to be completed in 2 weeks |
34 |
|
35 |
We have a limited number of slots (like every year) and we want to |
36 |
focus on the high value projects. I'm not saying this project is not |
37 |
interesting and people can work on whatever they want in their free |
38 |
time; but Google is paying for features here, so we will value |
39 |
projects that deal with features people will actually need and use. |
40 |
|
41 |
> |
42 |
> Marcelo |
43 |
> |
44 |
> (i did try to answer the original thread, but this is a new |
45 |
> subscription and the get archived mail action did not work) |
46 |
> |
47 |
> |