1 |
On 27/06/2022 21.39, Georgy Yakovlev wrote: |
2 |
> On Mon, 2022-06-27 at 21:21 +0200, Florian Schmaus wrote: |
3 |
>> Thanks to Mike Gilbert (floppym) for valuable feedback. |
4 |
>> |
5 |
>> Closes: https://bugs.gentoo.org/853928 |
6 |
>> Closes: https://github.com/gentoo/gentoo/pull/26069 |
7 |
>> Signed-off-by: Florian Schmaus <flow@g.o> |
8 |
>> --- |
9 |
>> eclass/java-vm-2.eclass | 30 ++++++++++++++++++++++++++++-- |
10 |
>> 1 file changed, 28 insertions(+), 2 deletions(-) |
11 |
>> |
12 |
>> diff --git a/eclass/java-vm-2.eclass b/eclass/java-vm-2.eclass |
13 |
>> index 8196b1cdc72a..dc0d87f4caf5 100644 |
14 |
>> --- a/eclass/java-vm-2.eclass |
15 |
>> +++ b/eclass/java-vm-2.eclass |
16 |
>> @@ -25,6 +25,7 @@ RDEPEND=" |
17 |
>> " |
18 |
>> DEPEND="${RDEPEND}" |
19 |
>> BDEPEND="app-arch/unzip" |
20 |
>> +IDEPEND="app-eselect/eselect-java" |
21 |
> |
22 |
> IDEPEND here will not do anything to current jdk source ebuilds. |
23 |
|
24 |
Right, this is specified providently. I've updated the commit message of |
25 |
the change in the according pull request to make this more clear. |
26 |
|
27 |
Also note that PMS does not guarantee that IDEPEND dependencies will be |
28 |
available in pkg_*rm phases. |
29 |
|
30 |
But ultimately, declaring IDEPEND on eselect-java is the best thing we |
31 |
can do here. |
32 |
|
33 |
|
34 |
> openjdk source ebuilds are EAPI7 at the moment and can't be updated to |
35 |
> EAPI8 due to econf unconditionally passing '--disable-static' |
36 |
|
37 |
I remember running into this when I tried to bump the JDK ebuilds to EAPI 8. |
38 |
|
39 |
- Flow |