Unrelated to anything: "Pull request" has always been a bad name. Always. They should have been named "Push requests".

@mcc I feel “Merge Requests” (what GitLab calls them iirc) was also a better name

@mcc I am fond of "Merge Request".

@mcc a place i used to work at used "merge requests" as the verbage. that seemed to make more sense

@mcc change proposal !

@mcc Agreed, cosigned, endorsed, and ratified.

@mcc Thank you!

I'm used to them now, of course, but when Git[Hub] was just taking off I assumed for the longest time that, like, you had to have a server the project maintainer could reach and that a "pull request" was asking them to thus do a pull from your server. And therefore that if "pull requests" were how you contributed to a project, then if you didn't have a setup like this you should just get lost, because that must be a sign that you aren't Serious Enough about open source software.

@klara @mcc as far as I understand it, this _was_ the original workflow, used by the kernel devs

of course, nobody else has ever done this

@r @klara doesn't lkml use an email based patch workflow?

@mcc In the place where they originated the name made sense, and then GitHub ruined it.

@mcc @r @klara usually, but for larger changes they use pull requests over email. see man git-request-pull

@mcc @klara i don't fully understand it tbh, it's always been too intimidating

but afaik the core subsystem maintainers use pull requests, whereas the rest of the work happens via email?

@r @klara oh no

@mcc in a painfully literal sense you have to request an action from someone else, so that action can only be "pulling." It's very confusing semantics.

@r @mcc Mm, so I was merely assuming that all open source projects were as inaccessible to contributions from casuals as the Linux kernel.

Interesting that I somehow got that impression.

@eramdam @mcc you just like it because saying their acronym makes you feel more like a cat

@klara Linux kernel development is accessible to casual contributions. My first contribution looked like this, in an e-mail to the Linux Kernel Mailing List: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=1c594c05a75770ab53a329fc4eb99c797a4bc7d7

@r @mcc

@zkat @mcc LMAO. I don't use GitLab so I never considered that let alone said "MR" out loud

@zkat @eramdam i was thinking it but i didn't want to say it

@wonka @klara @mcc it's not so much about getting individual patches merged (i could probably also do that if i really wanted to)

it's about the _public perception_, where even someone like me (who has "20 years of experience") generally feels as if it is too much effort to try and upstream patches (i.e. the _perceived/predicted_ effort isn't balanced by the reward/benefit)

@mcc @zkat @eramdam cats both PR and MR, I think he is unbiased in this matter.

@atonal440 @mcc @eramdam I... you know, you make a compelling argument. But we could really get into whether PR or MR is a cuter sound in general.

@mcc @eramdam one of us had to have the courage.

@mcc
I'd bet it was fine until they added the merge button

@kstatz12 @mcc Another long-standing term is "change request" (well from memory anyway, I didn't try very hard to confirm which systems use/used this wording)

@mcc Still not as bad as the X Window System making the shared machine the “client” and the desktop machine the “server”

@soaproot @mcc i think we avoided that due to the internal compliance form to change production data was known as a Change Request or CR for short.

@zkat @eramdam @mcc but cats also like to push things

@mcc It's a request to pull in some code you're trying to push. You're not requesting to push it, that is already happening 🤷

@jens but I could also frame it as "I'm requesting permission to push this code"

@mcc @jens yes but git doesn’t really have permissions does it? There is no way in git for anyone to fulfill the request for permission for you to push. But there is a way in git to pull. Both are reasonable framings but one requires things beyond what git affords for.

Don’t mean to argue. I am actually very bullish on reifying permissions and standardizing serialization of them so that we can request/receive/revoke them. I like your framing but it seems beyond git iiuc

@mcc then it would be a push permission request 😁

I don't deeply care about the naming, mind you, just explaining my understanding.

@mcc Other than grammar stuff, I think it also describes the mutual agreement. You offer your code, but the repo maintainer accepts it, so the final part is the pull. Well, just before the merge.

Merge requests are way more apt, FWIW.

@bengo @jens I mean either pull or push could be argued to be bad because they use git terminology and GitHub should be accessible to non git users.

@mcc Something git related with a bad name? 😏

@bagder @mcc
Yes, let‘s have a naming discussion first!😂

@icing @mcc I think we should call it a title discussion😎

@bagder @icing @mcc surely you're not changing things without setting up a working group with a standards body first

@icing @bagder @mcc To add to the naming discussion, I like "merge request" better, like "please merge this code I wrote".

@mijndert @icing @bagder "Alex, has there been any movement on my PMTCIW?"