| 1 |
Am Sonntag, 7. Mai 2017, 22:30:51 CEST schrieb Michał Górny: |
| 2 |
> |
| 3 |
> Oh, one more thing that I've forgot to mention in the original mail. |
| 4 |
> It'd probably be useful to solve two disjoint problems: |
| 5 |
> |
| 6 |
> a. whether CI should enforce correct depgraph and how, |
| 7 |
> |
| 8 |
|
| 9 |
Well, my approach for this would be that CI should enforce the same things as |
| 10 |
Repoman. |
| 11 |
|
| 12 |
Which means |
| 13 |
* stable: CI enforces separate deptrees for arch and ~arch |
| 14 |
* (testing, now) mixed: CI enforces deptree for ~arch (treating arch as ~arch) |
| 15 |
* unstable: CI enforces deptree for ~arch, CI errors on encountering arch |
| 16 |
|
| 17 |
> |
| 18 |
> 2. stable depgraph is broken and we don't care (i.e. we don't CC |
| 19 |
> the arch and drop old stable versions if necessary). |
| 20 |
|
| 21 |
Keep the arch on "mixed", or drop all stable keywords in one commit and then |
| 22 |
set it to "unstable". |
| 23 |
|
| 24 |
|
| 25 |
> |
| 26 |
> b. whether we should request stabilizing packages. |
| 27 |
> |
| 28 |
> |
| 29 |
> 1. stable depgraph is broken but we want to fix it (i.e. should CC |
| 30 |
> the arch on stablereqs), |
| 31 |
|
| 32 |
Well... so far my approach would have been, |
| 33 |
* keep arch (testing) mixed, until the arch team says "good to go", |
| 34 |
* then make hard switch to stable, and start with stablerequests |
| 35 |
|
| 36 |
We could, however, add one more column *only* for the (testing) mixed case, |
| 37 |
which states whether stabilization requests are required. |
| 38 |
|
| 39 |
(Only for the mixed case, since for "stable" this is always yes, and for |
| 40 |
"unstable" it makes no sense.) |
| 41 |
|
| 42 |
|
| 43 |
|
| 44 |
|
| 45 |
-- |
| 46 |
Andreas K. Hüttel |
| 47 |
dilfridge@g.o |
| 48 |
Gentoo Linux developer (council, perl, libreoffice) |
Archives