Re: [arch-general] [arch-dev-public] signoffs are dead
On Sat, 2016-07-02 at 09:16 +1000, Allan McRae wrote:
This sounds like the Fedora policy where packages have to surpass a certain karma level to move into the main repositories. I'm not sure who gets to vote for that though.
A
I would suggest allowing any Arch user vote yay or nay (with comments) - I believe this is what fedora did - only requirement was to be a registered user of the website. I don't imagine significantly more information will be obtained beyond the issues raised now in the mail list and forum. However, it does have the advantage of putting it in a single place for the packager. gene
As Florian just sent out the arch-dev-public email on this topic (and I can't reply there), I have a couple of questions. 1. What exactly can these users do? I assume simply sign-off as 'works for me' 2. Is there any hard deadline? 3. (related to above) is there any policy as to what specifically a user should have tested before signing off? Starting at least one binary (or all binaries?) etc.? On Tue, Jul 5, 2016 at 8:11 PM, Genes Lists via arch-general <arch-general@archlinux.org> wrote:
On Sat, 2016-07-02 at 09:16 +1000, Allan McRae wrote:
This sounds like the Fedora policy where packages have to surpass a certain karma level to move into the main repositories. I'm not sure who gets to vote for that though.
A
I would suggest allowing any Arch user vote yay or nay (with comments) - I believe this is what fedora did - only requirement was to be a registered user of the website. I don't imagine significantly more information will be obtained beyond the issues raised now in the mail list and forum. However, it does have the advantage of putting it in a single place for the packager. gene
On 26.07.2016 03:23, Oon-Ee Ng via arch-general wrote:
1. What exactly can these users do? I assume simply sign-off as 'works for me' 2. Is there any hard deadline? 3. (related to above) is there any policy as to what specifically a user should have tested before signing off? Starting at least one binary (or all binaries?) etc.?
It's really rather simple, at least for now. Maybe this changes when we have a few testers and can rely on getting signoffs quickly. As you mentioned, a signoff just says "I've done some basic testing and it works". In most cases that's just starting the main binary of the package and checking if it can deal with something. For example, reinstall some package with a new pacman, boot a new kernel, list the contents of some tarball with a new tar or run a simple script with a new perl version. There are no hard deadlines. If a package doesn't get enough signoffs (currently the limit is 2) within a reasonable time, it will just be moved without them. The "reasonable time" depends on the change and on what the maintainer feels to be sufficient. It could be anything from a few days for a simple security patch to a couple weeks for a new major release. There's also an old wiki page[1] which I've freshened up a bit. [1] https://wiki.archlinux.org/index.php/DeveloperWiki:CoreSignoffs PS: Please don't top post. Just don't quote anything if you don't need a quote. Florian
participants (3)
-
Florian Pritz
-
Genes Lists
-
Oon-Ee Ng