1 |
Christian Plessl wrote: |
2 |
|
3 |
> |
4 |
> If performance of portage is a problem for you, I would suggest using the |
5 |
> current portage information to build a fast searchable index for the |
6 |
> portage information. |
7 |
|
8 |
I agree with you. I think it should be better to have separate index |
9 |
which can be rebuild from scratch upon corruption or whatever. |
10 |
If only the searching speed is needed then only gentoolkit's qpkg |
11 |
could be changed, reducing the possibility of breaking portage |
12 |
However an extension to portage itself could be an option if we want |
13 |
alse speed up the rependency resolution (expecially for emerge -up |
14 |
world) or for future features like auto-useflag-change-detection. |
15 |
Portage's access to package metainformation could be separated in |
16 |
front and backend and modularized with failback on the text-files backend. |
17 |
The backends could be for text, berkleydb (little library dependency, |
18 |
useful on stage1 files or livecds), relational db or some other |
19 |
backstore. The frontends could include portage itself, a xml extractor |
20 |
(which could be an interface for other programs),... |
21 |
|
22 |
|
23 |
Marko |
24 |
|
25 |
|
26 |
-- |
27 |
gentoo-dev@g.o mailing list |