People, trim your damn email replies. Why do I need to scroll down hundreds of lines to find your acked/reviewed-by? Guess what, that needs ZERO context, as it's readily apparent which email you're replying to.

Or scoll down hundreds of lines to find a comment on a hunk. Trim the rest! And then scrolling another hundred lines just to find the rest is quoted text with zero replies.

It's just too lazy, and it wastes everybody's time. Trim!

@axboe

@amackif 🤦‍♂️

@axboe that was part of my motivation for writing this thing: https://mehlbrei.thejh.net/thread?msgid=cover.1750065793.git.asml.silence%40gmail.com#topic-0
(though to be clear this is a hacky prototype)

@jann nifty! Doesn't solve the time waste for "normal" clients though, nor the waste in general in terms of delivering all these emails that are 99% waste :-(

@axboe https://subspace.kernel.org/etiquette.html#trim-your-quotes-when-replying

@monsieuricon now if only people actually followed this, I would not need to be a grumpy old man about it...

I can understand that, and I do that in general for *comments*, but for the specific case of `Reviewed-by` and similar, some part of me thinks of those like signing something, and wants to quote exactly what I'm "signing". (That said, what I normally do is reply and insert my `Reviewed-by` right after the `Signed-off-by` in the commit message, not below the patch; I just leave the patch in below that. So, not much scrolling required.)

@josh but the email headers already very helpfully have a unique-id and your reply will show what it's replying too, so that really should not be necessary.

I think this is mostly a mailer problem, the people doing these spammy replies don't see the issue on their end. At least I hope it is, otherwise it's just laziness and lack of respect for other peoples time...

I do get your point, though.

@ljs @monsieuricon

> tbh people have different approaches to this.

They do - there's the wrong way, and then there's the right way. The approach chosen may be one of those two! Or in between.

The context is readily available in what is being replied to, or on lore. There's no excuse for not trimming. The cases of overly trimming are few and far between, compared to the wasteful cases of "insert single line in the middle of email, hit send".

Yeah, I get that. And it *does* bother me when I have to scroll through 50 uncommented patch hunks to find any responses; that's a pain, and doesn't seem like it serves any purpose.

@axboe @josh

Gmail not only hides the quotes, but actively makes it difficult to trim in the old-school way. This battle is lost, I'm afraid.

@mattdm @josh yeah, as usual it's (mostly) a client issue. Similarly to Microsoft introducing top replies, which has doomed us ever since.

@axboe I agree overall. But it's often not used in good faith. Too often it is used to ignore arguments or objections. So it becomes a tool for passive aggressive review.

This has knock-on effects because other people will jump in on the reply at which point the context is completely lost.

Then you end up having to either link or copy context back in. So I use it judiciously if possible.

@brauner The main issue is really that email sucks for this, period. It's a terrible tool.

But trimming at least makes it better.

And yes, there's misuse, like both you and Lorenzo point out. But I think both of those are corner cases compared to the main issue.

@axboe Meh, I don't find email that bad. I hate GitHub review equally. What I really love about b4 is that you can just pull everything into a single mbox and then go to town on it with whatever tool you like. That's why I keep urging @jann to make his thing work. 😅

@brauner @jann It is bad in the sense that there's no uniform way to handle it - which was my initial complaint here. I still greatly prefer email to GitHub or similar web based review, those all suck even more.

But the fact that existing tools suck doesn't mean that email is good, just that it's better than web based review. Which is a (very) low bar, imho.

@axboe @jann I don't think we'll ever get a Rosetta Stone moment for kernel code reviews. Ffs, we can't even agree on a unified way of testing things. 😅

@axboe @jann I think git is the closest we've ever gotten to a unified development tool.

@brauner @jann So what you're saying here is that we need Linus to make a new communication tool for us. Top down enforcement is the only way.

@axboe @jann You said it, not me.

@axboe trim, storage pun?

@axboe im of the opinion that any rule that needs to be followed should be codified

maybe a bot can auto reject emails containing tags and responses?

in any case this would prob be easier to setup with a code forge rather than email, but even with email it should be possible

@markcallaghan I’m discarding that comment

@ljs @axboe @monsieuricon I think in most cases it's actually possible to trim in a review of a long patch with some comments in random places.

A typical example of this sort would have, say, 30 hunks, and I have comments on five of them. Maybe I need to keep three more hunks for context so that readers can understand my comment on the following hunk. This means I can get rid of 22 hunks.

Or say it's just one very long hunk introducing a new file. I won't quote functions that get no comments.

@axboe Cannot agree more. Also complained and now I just ignore several emails. If sender does not have time to trim the huge email, I don't have time to read it (thus also consider their Rb/Ack).

@axboe I remember those good old days where not triming or top posting would get you sent to /dev/null. And now my OCD is triggered anytime I see those thousands of quoted lines just for one more line of text. And part of me is also triggered because of the lack of efficiency of all this.

@mathieu Indeed, tolerance used to be much lower. I wonder if that went away with the "gotta be nicer on the mailing list" times.

@axboe Yeah I think it is. Usually one wouldn't voice this opinion but I think it is. While the "nicer" part is valuable on some points, the low tolerance aspect was a good part of it imho.

On the other side when I speak like that my significant others tell me I'm an old fart :D

@monsieuricon @axboe Interesting, I didn't know about subspace.kernel.org. Is that what replace vger? I guess I've just been visiting lore and hadn't noticed.

@pdp7 @axboe Yes, subspace is where vger lives now.

@monsieuricon @axboe I'm suddenly self-conconsious today when replying to patches. I suppose I am guilty of this sometimes. /me grabs the ✂️