| 1 |
Right now repoman is relatively permissive - it whines about many things, but |
| 2 |
treats many issues as warning. |
| 3 |
The result is that many ebuilds get committed with 'minor' cosmetic issues |
| 4 |
which then someone more OCD than the original committer cleans up, making |
| 5 |
pretty much everyone involved more unhappy. |
| 6 |
|
| 7 |
There's no reason to not error out on, for example, an invalid RESTRICT. Just |
| 8 |
printing a message is relatively useless. |
| 9 |
|
| 10 |
|
| 11 |
Thus I suggest making the following warnings proper errors: |
| 12 |
|
| 13 |
(Taken from current repoman 'qawarnings' set) |
| 14 |
|
| 15 |
"changelog.missing", |
| 16 |
"changelog.notadded", |
| 17 |
"digest.assumed", |
| 18 |
"digest.unused", |
| 19 |
"ebuild.notadded", |
| 20 |
"ebuild.nesteddie", |
| 21 |
"DESCRIPTION.toolong", |
| 22 |
"RESTRICT.invalid", |
| 23 |
"ebuild.minorsyn", |
| 24 |
"ebuild.badheader", |
| 25 |
"metadata.warning", |
| 26 |
"LIVEVCS.stable", |
| 27 |
"LIVEVCS.unmasked", |
| 28 |
|
| 29 |
Most of these have few or no occurrences in the current tree, so changing the |
| 30 |
default from warn to error won't get in the way of the normal workflow. |
| 31 |
|
| 32 |
(A few of them, like DESCRIPTION.toolong, still have about a dozen leftovers, |
| 33 |
but that should be easy to fix) |
| 34 |
|
| 35 |
Have fun, |
| 36 |
|
| 37 |
Patrick |
Archives