| 1 |
-----BEGIN PGP SIGNED MESSAGE----- |
| 2 |
Hash: SHA1 |
| 3 |
|
| 4 |
Attaching a patch for java-ant-2. |
| 5 |
- - puts dev-java/ant-core into DEPEND, unless |
| 6 |
JAVA_ANT_DISABLE_ANT_CORE_DEP is defined not empty before inheriting |
| 7 |
java-ant-2, or java-pkg-opt-2 is inherited (we don't want ebuilds with |
| 8 |
optional java support hard-dep on ant) |
| 9 |
- - because of this, now java-ant-2 MUST be inherited after java-pkg-2 or |
| 10 |
java-pkg-opt-2. There's a check that dies if not, to enforce it. It may |
| 11 |
be a bit overkill (we could just check in java-pkg-opt-2 that java-ant-2 |
| 12 |
isn't already inherited) but I think it's more consistent and future |
| 13 |
proof to enforce the order for java-pkg-2 too. And inheriting just |
| 14 |
java-ant-2 with none of java-pkg-* makes no sense. |
| 15 |
|
| 16 |
I've updated the few cases where java-ant-2 was inherited before |
| 17 |
java-pkg-* in the tree and our overlays, so there should be no breakage. |
| 18 |
Please make sure any new ebuilds of yours have correct order. |
| 19 |
|
| 20 |
Hm now I realized that people who have these packages installed before |
| 21 |
update will get it die when unmerging. So enforcing the order like this |
| 22 |
is not possible. Maybe it could just ewarn instead of die? People will |
| 23 |
be able to unmerge, and devs should notice the complaint when working on |
| 24 |
ebuild if they get the order wrong. Or maybe we could check for ebuild |
| 25 |
phase? |
| 26 |
- -- |
| 27 |
Vlastimil Babka (Caster) |
| 28 |
Gentoo/Java |
| 29 |
-----BEGIN PGP SIGNATURE----- |
| 30 |
Version: GnuPG v1.4.6 (GNU/Linux) |
| 31 |
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org |
| 32 |
|
| 33 |
iD8DBQFFvqFFtbrAj05h3oQRAlkIAJ4+3pmoz9G/39+pQs55PlD536WUlQCfXzD2 |
| 34 |
n6d7/VReaDZEJxL+Mt3RJfk= |
| 35 |
=Q6Xk |
| 36 |
-----END PGP SIGNATURE----- |
Archives