1 |
On Thu, May 1, 2008 at 3:11 PM, <reader@×××××××.com> wrote: |
2 |
> In the middle of doing a major upgrade from very old pkgs to current |
3 |
> 2008 and compiling lots and lots of stuff. |
4 |
> |
5 |
> Seeing that line `checking for WHATEVER' go by 486,211 times so far |
6 |
> makes me wonder if there wouldn't be someway to cache all those |
7 |
> answers somewhere so whatever test is done for each line could be |
8 |
> dispensed with for most of them. Probably would need more than 2-3 |
9 |
> compiles to have all but rare ones answered. |
10 |
> |
11 |
> Some items really check a lot of things. |
12 |
> |
13 |
> I think it would be a major time saver when discussing huge numbers |
14 |
> of compiles. |
15 |
> |
16 |
> |
17 |
> -- |
18 |
> gentoo-user@l.g.o mailing list |
19 |
|
20 |
I had thought the same thing myself some time ago, and I discovered |
21 |
that there had been work on a FEATURE called confcache. I believe it |
22 |
was abandoned, though, due to major difficulties. This is merely a |
23 |
guess, but I think some of the problems arise in that some of the |
24 |
things that are checked for actually change as a package is installed |
25 |
or updated (e.g. checking gcc version). This means that each package |
26 |
being installed would have to somehow flag confcache and indicate that |
27 |
it has changed, and confcache would have to keep a list of all these |
28 |
cached values and their dependencies. |
29 |
|
30 |
I think there might be potential, however, for something that cached |
31 |
some of the more common system checks such as number of command line |
32 |
arguments. Then again, if many of these configuration items are |
33 |
discovered through a simple system call or by running a quick command, |
34 |
I'm not sure how much faster something like confcache would actually |
35 |
be. |
36 |
-- |
37 |
gentoo-user@l.g.o mailing list |