| 1 |
I have been looking at qa of the 2 java overlays (java-overlay and |
| 2 |
java-experimental) and have noticed a few things |
| 3 |
|
| 4 |
1) Running repoman ( and even qualidus, tho I still need to investigate |
| 5 |
this more ) on java-experimental is a waste of time. Both assume only |
| 6 |
that gentoo and java-experimental exist. This results in |
| 7 |
a) the inability to find eclasses, aka, java-maven-*.eclass; and |
| 8 |
b) the inability to find packages that are only located within java-overlay |
| 9 |
|
| 10 |
I have had a short discussion with zmedico about added the concept of a |
| 11 |
"parent" overlay so that repoman can find packages. |
| 12 |
So far zmedico has suggested merging java-overlay and java-experimental. |
| 13 |
|
| 14 |
2) QA within java-overlay is pretty poor. |
| 15 |
|
| 16 |
eg. |
| 17 |
KEYWORDS.stupid 5 |
| 18 |
dev-java/abbot/abbot-0.13.1-r1.ebuild |
| 19 |
dev-java/byaccj/byaccj-1.14.ebuild |
| 20 |
dev-java/gnu-crypto/gnu-crypto-2.1.0.ebuild |
| 21 |
dev-java/jdic/jdic-0.9.1.ebuild (believe this may have been my |
| 22 |
doing, certainly remember looking at this ) |
| 23 |
dev-java/libeds-java/libeds-java-0.5.1.ebuild |
| 24 |
|
| 25 |
There also seem to be a lot of packages that have *DEPENDS on packages |
| 26 |
within java-experimental, that are keyworded for arch's that don't have |
| 27 |
a package also keyworded for that arch. |
| 28 |
|
| 29 |
Now im going to start looking at all of these issues. But please ensure |
| 30 |
you run repoman and correct any issues you find. I also believe that we |
| 31 |
should always consider whether a change within the gentoo repo will have |
| 32 |
effects on java-overlay. While java-overlay is ~arch by nature lets not |
| 33 |
break it thru carelessness. |
| 34 |
|
| 35 |
Alistair. |
Archives