1 |
George Prowse wrote: |
2 |
> Caleb Cushing wrote: |
3 |
>> how about providing me a list of packages that are masked instead of |
4 |
>> making me read and unmask them one at a time. |
5 |
> That pretty much defeats the object of them being masked in the first place |
6 |
> |
7 |
> So all you can really come up with is speed? If a power user yourself |
8 |
> can only come up with speed what is an ordinary user going to think |
9 |
> of... *sigh* |
10 |
|
11 |
Evidently, you've never had to unmask a large package and test it. This |
12 |
is a major complaint I have with Portage. The reason I haven't brought |
13 |
it up however, is that I've been waiting until I had the time to have a |
14 |
look at the code behind `emerge` and see if I can embed the necessary |
15 |
logic to make this happen. |
16 |
|
17 |
Try `emerge php` on mips... Yeah, at some point it might be worth |
18 |
re-keywording it, but right now, one must go though and re-run emerge |
19 |
about 20 times as each unkeyworded dependency is identified. On a |
20 |
modern x86/AMD64 machine, that doesn't take too long, it's just an |
21 |
annoyance. On a slow Cobalt server however, it's agonising. |
22 |
|
23 |
KDE is another which I'd like to point out. I've got much of KDE 3.5.6 |
24 |
working on Loongson (at the moment, gpgme failing its tests holds me up) |
25 |
however, this didn't come without much arguing with portage over missing |
26 |
keywords. This is for monolithic KDE... I don't even want to think |
27 |
about the meta-ebuilds for KDE. |
28 |
|
29 |
Portage itself, does take a long time to load its libraries. I'm yet to |
30 |
try Paludis (I'm one of the few MIPS devs that doesn't use it yet), but |
31 |
do hear good things about it. Perhaps some of the concepts used in |
32 |
Paludis could be applied in portage, in order to speed up performance |
33 |
when searching or querying? |
34 |
|
35 |
These are again, minor changes, but improvements nonetheless. To say we |
36 |
haven't done anything in the last few years, would be a mistake in my |
37 |
opinion -- a lot has changed behind the scenes that users may not |
38 |
necessarily be aware of. I do not see this as being a significant |
39 |
problem. Certainly, the focus should be on the quality of the |
40 |
distribution; not quantity of features. |
41 |
|
42 |
Regards, |
43 |
-- |
44 |
Stuart Longland (aka Redhatter) .'''. |
45 |
Gentoo Linux/MIPS Cobalt and Docs Developer '.'` : |
46 |
. . . . . . . . . . . . . . . . . . . . . . .'.' |
47 |
http://dev.gentoo.org/~redhatter :.' |
48 |
|
49 |
I haven't lost my mind... |
50 |
...it's backed up on a tape somewhere. |