1 |
On Sun, 08 Mar 2009 08:49:16 +0100 |
2 |
Tiziano Müller <dev-zero@g.o> wrote: |
3 |
> With eapis 1 and 2 we introduced nice features but also a couple of |
4 |
> new problems. One of them are the use dependencies when the package |
5 |
> you depend on doesn't have the use flag anymore (see [1] for an |
6 |
> example). |
7 |
|
8 |
Here's another one to consider: |
9 |
|
10 |
If S= is wrong (which it often is, for packages with icky tarballs), |
11 |
src_configure and src_compile won't error out. These days this isn't a |
12 |
huge deal, because your custom src_install will probably fail. But if |
13 |
we're introducing a default src_install, it will instead see no |
14 |
Makefile and just do nothing, resulting in an empty package being |
15 |
installed. |
16 |
|
17 |
Currently, at the start of src_configure, the package manager does a cd |
18 |
to ${S} if ${S} exists, and to ${WORKDIR} otherwise. I'd like to |
19 |
propose that ${S} not existing should instead be an error if either of |
20 |
the following conditions are met: |
21 |
|
22 |
* ${A} is non-empty |
23 |
* Any of src_unpack, src_configure, src_compile or src_install is a |
24 |
defined phase. |
25 |
|
26 |
Ebuilds where this would trigger a false positive would have to specify |
27 |
S=${WORKDIR}. |
28 |
|
29 |
-- |
30 |
Ciaran McCreesh |