| 1 |
On 10/07/17 20:53, Rich Freeman wrote: |
| 2 |
> On Mon, Jul 10, 2017 at 3:09 PM, Matt Turner <mattst88@g.o> wrote: |
| 3 |
>> For what it's worth, Jack Morgan was recently getting his sparc and |
| 4 |
>> ia64 systems back up, but then decided to retire instead when he saw |
| 5 |
>> all of the discussions about dropping the architectures he cares |
| 6 |
>> about. |
| 7 |
>> |
| 8 |
> Honestly, I don't really get this sort of thing. The reason arches |
| 9 |
> get dropped is because they mark things stable that they can't keep up |
| 10 |
> with. If an arch never marked a package as stable nobody would be |
| 11 |
> bothered. If they only marked a few critical packages as stable and |
| 12 |
> then kept up with them, again nobody would be bothered. The conflict |
| 13 |
> comes in when an arch team marks packages as stable, but then doesn't |
| 14 |
> keep up with them. |
| 15 |
> |
| 16 |
> Marking a package as stable is a two-way commitment. When an arch |
| 17 |
> team marks a package as stable they make a promise to the maintainer |
| 18 |
> to stabilize updates in a timely manner. In return the maintainer |
| 19 |
> promises to keep older versions around to suit the needs of the arch |
| 20 |
> team for the short time it takes to do these stabilizations. |
| 21 |
> |
| 22 |
> When an arch team stabilizes something that they don't have time to |
| 23 |
> maintain then they're making a promise they can't keep, and the deal |
| 24 |
> breaks down. Eventually the maintainers complain, and the council |
| 25 |
> ends up revoking the right of the arch team to hold the maintainers to |
| 26 |
> their side of the deal which has already been broken. |
| 27 |
> |
| 28 |
> There are no bad guys here. There is just a certain amount of work it |
| 29 |
> takes to make a stable arch viable, and it either happens or it |
| 30 |
> doesn't. Most people who use Gentoo are tinkerers by nature. All |
| 31 |
> things being equal we'd love to see every arch supported. However, |
| 32 |
> this requires discipline on the part of the arch team, because |
| 33 |
> otherwise an arch that few people use starts impacting work for other |
| 34 |
> arches that many more use as maintainers get buried in old bugs. |
| 35 |
> |
| 36 |
I dunno where you've been lately, Rich, but for most devs, would-be |
| 37 |
devs, and observers .. there -are- no arch teams left .. just a few Arch |
| 38 |
devs, or arch 'people' .. |
| 39 |
|
| 40 |
This is why stabilisation, if not for individual package maintainers on |
| 41 |
amd64, has become a joke, save for Ago's efforts, and recent efforts by |
| 42 |
kensington to streamline the effort for the likes of ago with his bot, |
| 43 |
and one or two other arch stabilisers (who I know exist, but not by name |
| 44 |
or nick). |
| 45 |
|
| 46 |
There is no, and has not been, in the time I've been involved with |
| 47 |
Gentoo, any "pact" or "contract" between arch teams/devs and maintainers |
| 48 |
whatsoever, anything is only ever done as a 'favour' or if someone |
| 49 |
nudges the AT after the appropriate bug has been filed. |
| 50 |
|
| 51 |
Keywording, unless for your own native arch, is a worse joke .. you have |
| 52 |
to do it yourself .. there just aren't people with the less common |
| 53 |
arches actively seeking packages out to keyword. |
| 54 |
|
| 55 |
So, even for me, Thank You Ago for the packages you've kindly fixed |
| 56 |
where I have nudged you, and also maekke on the few arm packages where |
| 57 |
you have done the same. Your efforts are noted by some, pity not by others. |
Archives