| 1 |
On Fri, Aug 05, 2016 at 10:36:41AM -0400, Rich Freeman wrote: |
| 2 |
|
| 3 |
> No argument, but if you actually asked stable users I'm not convinced |
| 4 |
> that they'd prefer less-tested recent packages over well-tested older |
| 5 |
> ones. Anything to get things fresher is good, but there probably |
| 6 |
> needs to be some kind of sanity check. |
| 7 |
My two cents on the subject. Perhaps I am odd in this but in spite of |
| 8 |
having been messing with ebuild development I keep my Gentoo systems |
| 9 |
mostly stable - and I definitely prefer keeping it this way. This might |
| 10 |
perhaps have something to do with the fact I do not really need the |
| 11 |
latest and supposedly greatest versions of all packages, for instance |
| 12 |
the amd64 box I am writing this on has only got 15 entries (plus |
| 13 |
dependencies of those) in my "version bump" Portage keyword files. |
| 14 |
|
| 15 |
That said, the situation gets much worse for packages which can be found |
| 16 |
*only* in ~arch - on the same box the "not in stable" keyword file |
| 17 |
contains 72 entries. Only a few of these have been orphaned so I guess I |
| 18 |
could always submit a lot of STABLEREQs but telling the truth, the |
| 19 |
number has really discouraged me. |
| 20 |
|
| 21 |
Bottom line: I would say we do need some way of streamlining ebuild |
| 22 |
stabilisation. |
| 23 |
|
| 24 |
Cheers, |
| 25 |
-- |
| 26 |
MS |
Archives