1 |
2012/3/31 Tiziano Müller <dev-zero@g.o>: |
2 |
> Am Samstag, den 31.03.2012, 14:44 +0200 schrieb Ulrich Mueller: |
3 |
>> >>>>> On Sat, 31 Mar 2012, Maciej Grela wrote: |
4 |
>> |
5 |
>> > I've read the PMS and I haven't found information whether this variable |
6 |
>> > is supposed to be set during pkg_prepare or not. |
7 |
>> |
8 |
>> There is no such stage. You mean pkg_pretend, I suppose? |
9 |
>> |
10 |
>> > Therefore I ask, what is the proper behaviour here ? Is there |
11 |
>> > documentation on what special env variables are supposed to be |
12 |
>> > defined in each stage ? |
13 |
>> |
14 |
>> It's specified here: |
15 |
>> <http://dev.gentoo.org/~ulm/pms/4/pms.html#x1-11900012.1> |
16 |
>> |
17 |
>> | Variable Legal in Consistent? Description |
18 |
>> | --------------------------------------------------------------------- |
19 |
>> | T All Partially⁴ The full path to a temporary |
20 |
>> | directory for use by the ebuild. |
21 |
>> | |
22 |
>> | ⁴Consistent and preserved across a single connected sequence of |
23 |
>> | install or uninstall phases, but not between install and uninstall. |
24 |
>> | When reinstalling a package, this variable must have different |
25 |
>> | values for the install and the replacement. |
26 |
>> |
27 |
>> > Can this be considered as a bug in paludis ? |
28 |
>> |
29 |
>> The spec seems to be clear that T is legal in all phases, including |
30 |
>> pkg_pretend. |
31 |
> |
32 |
> Well, I'd say: there is no sane value you can assign to $T since you are |
33 |
> not allowed to write anything anyway: |
34 |
|
35 |
indeed it's not writing, check-reqs.eclass is reading space available |
36 |
via 'df -Pm' there is something else it could use assuming |
37 |
$PORTAGE_TMPDIR is worse for portability? |
38 |
|
39 |
> |
40 |
> "pkg_pretend must not write to the |
41 |
> filesystem." (http://dev.gentoo.org/~ulm/pms/4/pms.html#x1-9700010.1.2) |
42 |
> |
43 |
> and since "pkg_pretend is run separately from the main phase function |
44 |
> sequence, and does not participate in any kind of environment saving" it |
45 |
> is not guaranteed to be set to the same $T later. |
46 |
> |
47 |
> Cheers, |
48 |
> Tiziano |