On 9/29/22 13:53, Konstantin Ryabitsev wrote:
On Thu, Sep 29, 2022 at 01:31:49PM +0000, Artem S. Tashkinov wrote:
To me it sounds like the best way to keep moving forward is simply
convert git.kernel.org + patchwork.kernel.org + bugzilla to
gitlab.kernel.org and that will solve all the issues immediately.
No, you will still have all the exact same problems as long as nobody is in
charge of handling incoming bugs. There are plenty of active github/gitlab
projects with thousands of open issues that nobody is working on for the exact
same reason nobody is working on bugs filed via bugzilla -- the right people
didn't see it (or are actively ignoring it, because they are working on
something else).
That will require of course a ton of work but:
1) All the commiters will be automatically present and you can easily CC
them
Just like with bugzilla.
2) All the kernel directories could be split into components with the
respective developers being subscribed to them automatically. There's an
issue though: sometimes directories/components are rearranged. Gitlab
however is quite powerful, so issues can be easily moved between
components.
Just like bugzilla.
3) It's gonna be a ton easier to keep track of commits and discuss/review
them. AFAIK it's now done using LKML + patchwork.kernel.org and then
commits are merged by maintainers. So many places to keep track of.
Now there will be a single place someone can knock out to stop all kernel
development and coordination, subject to countrywide IP blocks, political
influence, etc.
Besides, maintainers already have a single place to keep track of things --
their inbox. If they didn't see something in their inbox, how is it going to
be different when it's a web interface?
4) Gitlab probably can be integrated with other gitlabs (at least AMD, Intel
and Nouveau drivers are developed on gitlab.freedesktop.org).
Gitlab simplifies all of that tremendously. Github will work as well but I
know many people don't like it.
Gitlab is also a commercial open-core project. It is permanently in danger of
being swallowed by some $ENTITY_NOBODY_LIKES, who will for sure look to
prioritize what kinds of things go in to the "open core" part and what kinds
of things are only available with subscription, in order to improve profit
margins.
-K
Not going to argue about that.
That leaves us with Bugzilla that no one wants to touch and some people
actively want to delete altogether. In other words, no central place to
report bugs or keep track of them.
I've mentioned several times already that mailing lists are _even worse_
in terms of reporting issues. Developers get emails and simply ignore
them (for a multitude of reasons).
And finding and engaging with existing issues becomes near impossible.
With Bugzilla you can easily leave a comment, attach a patch, attach
debug files, etc. For many mailing lists following up on an old message
is not possible unless you know the old message ID (the nature of mail
clients). With bugzilla you can see a list of reported/open bugs. With
bugzilla you can easily engage with other bug reporters.
To me it all sounds like kernel developers simply do not want any
responsibility or any extra emails whatsoever and instead would love
everyone to spam LKML: when you feel like it, you can check your inbox,
maybe leave a message, maybe not. Who cares? It's LKML.
Getting back to my first message in this discussion,
* Let's refresh all the components in Bugzilla
* Components may not have any people responsible for them at all. Bug
reporters will have to CC the people they are interested in.
* Let's subscribe the past six months of developers (using git commit logs)
* Whoever wants to unsubscribe is free to do so.
Alternatively:
* Delete all the components.
* Leave a catch-all one.
* Let bug reports rot because no one will ever see them. Almost just
like now. Don't remind me of mailing lists.
If not for bugzilla, let's use something more modern. I don't know any
comparable projects however. Trac is truly horrible. You cannot even
unsubscribe from bug reports. Maybe I've missed something. Discourse?
Not a bug tracker per se but can certainly work this way.
To me it all sounds like the sentiment is to absolve from any and all
responsibility and shut your ears and eyes to any reports of your code
malfunctioning/breaking. Fine, let's do it.
Sarcasm and pain aside, Linus Torvalds himself _via Bugzilla_ has helped
me resolve critical issues on several occasions while my messages to
LKML were simply _ignored_. Think about that.
Mailing lists will not work for such a huge project. Period. In the
early 90s they worked, but we are 25 years later with millions more
users. With a ton more of a ton more complicated hardware.
Maybe let's try Discourse with just a few "forums":
* arch
* drivers
* fs
* block/init/ipc/kernel/mm/crypto/virt/scripts - a single core component
* net
* security
* sound
* tools
That's it. Just 8 forums, zero maintenance, no need to add/remove/manage
anything.
Best regards,
Artem
