1 |
On Mon, Sep 8, 2008 at 1:58 AM, Santiago M. Mola <coldwind@g.o> wrote: |
2 |
> On Mon, Sep 8, 2008 at 10:44 AM, Alec Warner <antarus@g.o> wrote: |
3 |
>> |
4 |
>> Most obvious failure cases these days have build logs and the build |
5 |
>> logs will specify what the configure command |
6 |
>> was, so the only problematic area is looking at the ebuild to |
7 |
>> determine what will happen during execution. Arguably having |
8 |
>> an ebuildshell would assist here. However, ebuilds are already |
9 |
>> sufficiently complicated by eclass inheritance that reading |
10 |
>> many ebuilds is already difficult and I think this extension does not |
11 |
>> make that significantly worse. |
12 |
> |
13 |
> If you're just dealing with the default phases, it's not too hard to |
14 |
> say in advance the exact command that will be executed. Default phases |
15 |
> are well-defined in PMS. So you can look at them to see what will |
16 |
> happend if you define some variable. |
17 |
|
18 |
The point is that it is not in the ebuild; you need to look it |
19 |
up...unless you have all the default functions memorized? :) |
20 |
|
21 |
> |
22 |
> For example, for the proposed arguments for src_configure, a |
23 |
> definition would be something like this (taken from Exherbo, just |
24 |
> pretend it says USE instead OPTION and you're done): |
25 |
> default_src_configure() |
26 |
> { |
27 |
> if [[ -x ${ECONF_SOURCE:-.}/configure ]] ; then |
28 |
> econf \ |
29 |
> "${DEFAULT_SRC_CONFIGURE_PARAMS[@]}" \ |
30 |
> $(for s in "${DEFAULT_SRC_CONFIGURE_OPTION_ENABLES}" ; do \ |
31 |
> option_enable ${s} ; \ |
32 |
> done ) \ |
33 |
> $(for s in "${DEFAULT_SRC_CONFIGURE_OPTION_WITHS}" ; do \ |
34 |
> option_with ${s} ; \ |
35 |
> done ) |
36 |
> fi |
37 |
> } |
38 |
> |
39 |
> It's quite straightforward. |
40 |
> |
41 |
> Regards, |
42 |
> -- |
43 |
> Santiago M. Mola |
44 |
> Jabber ID: cooldwind@×××××.com |
45 |
> |
46 |
> |