1 |
On 04/06/2010 07:22 AM, James Cloos wrote: |
2 |
>>>>>> "ZM" == Zac Medico <zmedico@g.o> writes: |
3 |
> |
4 |
> ZM> You can configure eclass override behavior via eclass-overrides in |
5 |
> ZM> /etc/portage/repos.conf, as documented in `man portage`. |
6 |
> |
7 |
> ,----< From that manpage > |
8 |
> | When using eclass-overrides, due to bug #276264, you must ensure that |
9 |
> | your portage tree does not contain a metadata/cache/ directory. |
10 |
> `---- |
11 |
> |
12 |
> Which translates into "eclass-orderrides are completely and entirely |
13 |
> useless, so don't bother. |
14 |
|
15 |
Well, it's roughly equivalent to the old default behavior (which you |
16 |
apparently preferred). However, the issue is now complicated by the |
17 |
fact that FEATURES=metadata-transfer is disabled by default, so when |
18 |
portage goes to pull cache directly from metadata/cache/, it won't |
19 |
be able to validate eclass changes since there are no eclass |
20 |
timestamps saved inside metadata/cache/. FWIW, there was some |
21 |
discussion about extending the cache format to improve |
22 |
the validation mechanism for eclasses here: |
23 |
|
24 |
http://archives.gentoo.org/gentoo-dev/msg_cfa80e33ee5fa6f854120ddfb9b468b3.xml |
25 |
|
26 |
> Portage used to used to search for eclasses starting in the top overlay; |
27 |
> it should not have changed. |
28 |
|
29 |
Well, the biggest caveat to that behavior is that it tends to |
30 |
invalidate metadata cache as reported in this bug: |
31 |
|
32 |
http://bugs.gentoo.org/show_bug.cgi?id=124041 |
33 |
|
34 |
I'd be happy to work on resolving issues with eclass-orderrides to |
35 |
make it more usable. However, due to the need to regenerate metadata |
36 |
cache, I don't think that this is something that can ever again be |
37 |
enabled by default. |
38 |
-- |
39 |
Thanks, |
40 |
Zac |