On Mon, Mar 18, 2019, 04:31 Morten Linderud via arch-dev-public <arch-dev-public@archlinux.org <mailto:arch-dev-public@archlinux.org>> wrote: On Wed, Mar 13, 2019 at 02:05:47PM -1000, Gaetan Bisson via arch-dev-public wrote: > I think it's a great idea but it needs a solid maintainer. Without a > clear leader it's (probably) going to be a free for all and we'll drown > under bikeshedding issues within a month. But of course that doesn't > mean we'd lose anything trying anyhow. > > Among other things, I'd personally like to see the repo maintainer > enforce sensible and consistent naming for the tools, preferring longer, > explicit names over shorter ones. For instance, I'm sure many of us have > one-letter scripts and if we contribute them all there's bound to be > collisions along with the problem of not knowing at first glance what > each tool does. We could maintain a bash alias file containing > everyone's favorite nickname for each tool. If we want someone to a dedicated maintainer, I can probably do so. But I believe that we can give everyone commit access, block commits to master and just enforce a system where two reviews are needed before merge. I really don't think more is needed, but as noted; I can probably take some responsibilities if the devs think that is warranted. When it comes to packaging and naming conflicts, I wonder if it's just easier to drop all the supplied files into `/usr/share/archcontrib` or something. Makes it easier to package and doesn't clutter anyone's PATH with a lot of (sometimes) unneeded tools. -- Morten Linderud PGP: 9C02FF419FECBE16 -----BEGIN PGP SIGNATURE----- iQIzBAABCgAdFiEEktnGzemaICTWkKdu50JoO6CMsv8FAlyOvK0ACgkQ50JoO6CM sv8lBQ//aoNFKxUDM4Qcpi6nXXDDw51loxJp4AB+SYTv/Vml5tag7fVAkniUphLD ujJYY8/sSVqke29YYpG+OyJIIp91jP7WJ6nwtRVY2lsfOOUGkEX3HKEN0NCHxe9/ GpQHn1GqZAp3+FjSwxvJabpm+aGDL4CLVErdbiDKocpZJf61WSAXNHpGPj26PzUm L+fIkPDu8SEqENBVVkC9cKnf2+oInQ2ECHknSN/lBbNbwh2z2kAW46kglJXryJcE jO2gYt9kzU6NqcfKwHXHY0XCQd0H0pVMjfaK6PB1N+dwbgHAAbyhV/SIziFEP9kc PDK5NAFAcgQI+999Q9T0nufc+lqvPn9MuQ89OPINQhwTGCxEkUlicVrBYY/9ZFhC rrsTp7LExeBsMhAE6havnv2UHgkJH9ttv3FmSi81HsUYsgrxsUXpijqZUSacra5l 4VEGR9pPjsdERco2UZo9hcneLutQn0T2mCVLgIiCdjS2ZjsQuuHZ5RNBkyeNnWOC lBOBq7lddxhIiWGAjz60KOox5KL68OYlY4mZXbMQ2x2n3v2XF5VVL/ncp9F9CFo4 J6JGjab38LEemMnOFBgFS2XkRMmV7G+siD1++VdEKSCoy2tICPm5SAQiG9Ueo48y nUMD2+gfkzFNJmVQdym9YVxAKlbLWDGVHGd9/a43ZzjKKIReJgk= =y2os -----END PGP SIGNATURE----- So shall we progress with this? I don't think we can lose anything by just trying it out. Let's put up a GitHub repo and get coding. Of course it should be mirrored to our infra.