1 |
On Sunday 19 September 2004 21:33, Dan Armak wrote: |
2 |
> As I said, I'm definitely going to try making this work, but won't have |
3 |
> real time until the weekend after next (and most days after that, |
4 |
> hopefully), so you can beat me to it. |
5 |
|
6 |
There are other few issues, for example some libraries aren't installed in the |
7 |
system because defined in the Makefile.am as "noinst_LTLIBRARIES". And some |
8 |
programs in differents subdirs are linking to them, so these library must be |
9 |
recompiled for every program we are compiling that needs them and this will |
10 |
bring to more compilation time. I can't see a solution to this, installing |
11 |
them will go against the kde developers decisions and there's a reason |
12 |
because they aren't installed. |
13 |
|
14 |
|
15 |
> I've just tried the confcache patch. It's integrated into portage .51 cvs |
16 |
> snapshots here http://dev.gentoo.org/~ferringb/ebuild-daemon/51-pre20-cvs/ |
17 |
> - they apparently include other highly experimental stuff and the last one |
18 |
> has some bugs already fixed in cvs, so be in touch with #gentoo-portage if |
19 |
> using these. To enable confcache, add 'confcache' to FEATURES and change |
20 |
> the kde ebuilds to use econf rather than call configure directly. This |
21 |
> results in the kde.eclass attached - the change is trivial, but as I'm on |
22 |
> rsync not cvs atm, I can't make a diff immediately, sorry. The change for |
23 |
> split packages should also be on the order of a five-liner in kde.eclass, |
24 |
> apart from the makefile changes. |
25 |
> |
26 |
> On my athlon xp 2600, this makes the kdelibs configure run go down from 66 |
27 |
> seconds to 28 seconds. At least 10-12 seconds of each of these two numbers |
28 |
> is due to makefile generation, which will be very much reduced in split-up |
29 |
> ebuilds, so we get 54 sec -> 16 sec, or a 5x improvement. |
30 |
> |
31 |
> Also, on slower machines a far larger percent of time spent in (non-cached) |
32 |
> configure is due to slow compilations rather than (as here) IO, so the |
33 |
> benefit will be even greater there. This indicates we should at least |
34 |
> reevluate the emerge performance factor of splitting up the kde ebuilds. |
35 |
> |
36 |
> Using my old rough formula, we get: |
37 |
> 17 configure scripts (before splitting ebuilds) --> 20-60 minutes total |
38 |
> running time |
39 |
> splitting up --> >=220 packages (? but at least that many, I believe) |
40 |
> splitting up without confcache --> 220-660 minutes, i.e. 3.6-11 hours, |
41 |
> total running time (clearly unacceptable) |
42 |
> splitting up with confcache --> about 15-20% of above, i.e. 1-2.2 hours |
43 |
> total running time |
44 |
> |
45 |
> According to these numbers, if we both add confcache and split the ebuilds, |
46 |
> the total time for emerging all or nearly all of kde will increase by |
47 |
> 0.6-1.2 hours, plus the overhead from running the emerge cycle 200 more |
48 |
> times, which I hope is relatively negligible (a few minutes?). |
49 |
> |
50 |
> Compared with the benefits (to those who don't want all of kde, and |
51 |
> considering that >95% of people typing 'emerge kde' to save package |
52 |
> selection time don't really want kdetoys and kdeedu and such, this seems on |
53 |
> first sight to be a reasonable tradeoff for the added functionality. What |
54 |
> do you think, everyone? Also note that my benchmarking is hardly |
55 |
> scientific, so I'd be glad if someone bothered to repeat it and compare his |
56 |
> results to mine. |
57 |
I'll try it. Thanks. |
58 |
|
59 |
Bye! |
60 |
--- |
61 |
Simone Gotti |
62 |
<simone.gotti@×××××.it> |
63 |
http://kde-bluetooth.sf.net |
64 |
|
65 |
-- |
66 |
gentoo-dev@g.o mailing list |