KTODO automated TODO lists

KTODO automated TODO lists

[Dan Carpenter]

Yesterday someone on my lists just sent an email looking for kernel
tasks.  This was a university student in a kernel programming class.
We also have kernel-janitors and outreachy and those people are always
asking for small tasks.

One thing that I sometimes say as a reviewer is "This isn't caused by to
your patch but I noticed something wrong."  We could add that kind of
thing to a todo list by using a KTODO line.

KTODO: add check for failure in function_something()

Then people can look on lore or use lei to find small tasks to work on
or they could use lei.

lei q -I https://lore.kernel.org/all/ -o ~/Mail/KTODO --dedupe=mid 'KTODO AND rt:6.month.ago..'

Then grep ^KTODO ~/Mail/KTODO -R and cat the filename you want.

regards,
dan carpenter

Re: KTODO automated TODO lists

[Linus Walleij]

On Thu, Oct 19, 2023 at 6:11 AM Dan Carpenter <dan.carpenter@linaro.org> wrote:

> We could add that kind of
> thing to a todo list by using a KTODO line.
>
> KTODO: add check for failure in function_something()
>
> Then people can look on lore or use lei to find small tasks to work on
> or they could use lei.
>
> lei q -I https://lore.kernel.org/all/ -o ~/Mail/KTODO --dedupe=mid 'KTODO AND rt:6.month.ago..'
>
> Then grep ^KTODO ~/Mail/KTODO -R and cat the filename you want.

I like it! There are too many of these things falling on the floor.
An easy way to stash it on the technological debt hitlist would be
really helpful.

Yours,
Linus Walleij

Re: KTODO automated TODO lists

[Geert Uytterhoeven]

On Thu, Oct 19, 2023 at 2:50 PM Linus Walleij <linus.walleij@linaro.org> wrote:
> On Thu, Oct 19, 2023 at 6:11 AM Dan Carpenter <dan.carpenter@linaroorg> wrote:
> > We could add that kind of
> > thing to a todo list by using a KTODO line.
> >
> > KTODO: add check for failure in function_something()
> >
> > Then people can look on lore or use lei to find small tasks to work on
> > or they could use lei.
> >
> > lei q -I https://lore.kernel.org/all/ -o ~/Mail/KTODO --dedupe=mid 'KTODO AND rt:6.month.ago..'
> >
> > Then grep ^KTODO ~/Mail/KTODO -R and cat the filename you want.
>
> I like it! There are too many of these things falling on the floor.
> An easy way to stash it on the technological debt hitlist would be
> really helpful.

And if people use appropriate Closes: tags, someone can write a tool
to only list non-closed items.

Gr{oetje,eeting}s,

                        Geert


--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: KTODO automated TODO lists

[Liam R. Howlett]

* Geert Uytterhoeven <geert@linux-m68k.org> [231019 09:21]:
> On Thu, Oct 19, 2023 at 2:50 PM Linus Walleij <linus.walleij@linaroorg> wrote:
> > On Thu, Oct 19, 2023 at 6:11 AM Dan Carpenter <dan.carpenter@linaroorg> wrote:
> > > We could add that kind of
> > > thing to a todo list by using a KTODO line.
> > >
> > > KTODO: add check for failure in function_something()
> > >
> > > Then people can look on lore or use lei to find small tasks to work on
> > > or they could use lei.
> > >
> > > lei q -I https://lore.kernel.org/all/ -o ~/Mail/KTODO --dedupe=mid 'KTODO AND rt:6.month.ago..'
> > >
> > > Then grep ^KTODO ~/Mail/KTODO -R and cat the filename you want.
> >
> > I like it! There are too many of these things falling on the floor.
> > An easy way to stash it on the technological debt hitlist would be
> > really helpful.
> 
> And if people use appropriate Closes: tags, someone can write a tool
> to only list non-closed items.

Could the tool also look at potential git commits to the affected
function if Closes is lacking (potential opened issues)?  Sort of like
the stable kernel failed to apply patch emails list potential patches
that may also be needed?

We all try our best to have closes tags, etc but sometimes they are
missed.

Thanks,
Liam

RE: KTODO automated TODO lists

[Bird, Tim]

> -----Original Message-----
> From: Geert Uytterhoeven <geert@linux-m68k.org>
> On Thu, Oct 19, 2023 at 2:50 PM Linus Walleij <linus.walleij@linaroorg> wrote:
> > On Thu, Oct 19, 2023 at 6:11 AM Dan Carpenter <dan.carpenter@linaroorg> wrote:
> > > We could add that kind of
> > > thing to a todo list by using a KTODO line.
> > >
> > > KTODO: add check for failure in function_something()
> > >
> > > Then people can look on lore or use lei to find small tasks to work on
> > > or they could use lei.

You can also create dashboards for people to measure the amount of
desired work outstanding, and to easily find it.  I do something similar
with my own projects (but with a FIXTHIS: prefix)

> > >
> > > lei q -I https://lore.kernel.org/all/
> dKdTl7e3y1zlxphutZsrG0BoJ1YQRFA--MRHazjWXrH4QqVG0Hrp$ -o ~/Mail/KTODO --dedupe=mid 'KTODO AND rt:6.month.ago..'
> > >
> > > Then grep ^KTODO ~/Mail/KTODO -R and cat the filename you want.
> >
> > I like it! There are too many of these things falling on the floor.
> > An easy way to stash it on the technological debt hitlist would be
> > really helpful.
> 
> And if people use appropriate Closes: tags, someone can write a tool
> to only list non-closed items.

Wouldn't you remove the "KTODO:" and alter the comment (or remove it), with
the commit that closed the issue?  It seems like leaving the KTODO would
be a bug, as it no longer applies and would be confusing.

 -- Tim

Re: KTODO automated TODO lists

[Dan Carpenter]

On Thu, Oct 19, 2023 at 04:30:34PM +0000, Bird, Tim wrote:
> > And if people use appropriate Closes: tags, someone can write a tool
> > to only list non-closed items.
> 
> Wouldn't you remove the "KTODO:" and alter the comment (or remove it), with
> the commit that closed the issue?  It seems like leaving the KTODO would
> be a bug, as it no longer applies and would be confusing.

No, it's not a comment.  It's just an email to the list when you think
of something so you add a KTODO in the email.  It's like a searchable
hashtag, except it's a one line summary.

regards,
dan carpenter

RE: KTODO automated TODO lists

[Bird, Tim]

> -----Original Message-----
> From: Dan Carpenter <dan.carpenter@linaro.org>
> On Thu, Oct 19, 2023 at 04:30:34PM +0000, Bird, Tim wrote:
> > > And if people use appropriate Closes: tags, someone can write a tool
> > > to only list non-closed items.
> >
> > Wouldn't you remove the "KTODO:" and alter the comment (or remove it), with
> > the commit that closed the issue?  It seems like leaving the KTODO would
> > be a bug, as it no longer applies and would be confusing.
> 
> No, it's not a comment.  It's just an email to the list when you think
> of something so you add a KTODO in the email.  It's like a searchable
> hashtag, except it's a one line summary.

OK -  I misread the original proposal.  Sorry about that.
 -- Tim

Re: KTODO automated TODO lists

[Konstantin Ryabitsev]

On Thu, Oct 19, 2023 at 07:11:36AM +0300, Dan Carpenter wrote:
> KTODO: add check for failure in function_something()
> 
> Then people can look on lore or use lei to find small tasks to work on
> or they could use lei.

We can also have bugbot turn these into bugzilla bugs, if there's consensus
that it would be helpful.

-K

Re: KTODO automated TODO lists

[Andrew Morton]

On Thu, 19 Oct 2023 07:11:36 +0300 Dan Carpenter <dan.carpenter@linaro.org> wrote:

> Yesterday someone on my lists just sent an email looking for kernel
> tasks. 

Well here's a task: write a bot which follows the mailing lists and
sends people nastygrams if one of their emails is more than 95%(?)
quoted text.

It's happening significantly more lately.  Possibly because the gmail
client helpfully hides quoted text.

Probably not a great way of becoming popular.

Re: KTODO automated TODO lists

[Linus Torvalds]

On Mon, 23 Oct 2023 at 08:49, Andrew Morton <akpm@linux-foundation.org> wrote:
>
> Well here's a task: write a bot which follows the mailing lists and
> sends people nastygrams if one of their emails is more than 95%(?)
> quoted text.

I think that might be better off as a spam filter rule.

Don't make it some after-the-fact "trawl the lists". Just make it a
bounce with a "you quoted too much". Same as the html avoidance.

Make it ok to quote 15 lines of commit message for a "Reviewed-by:"
kind of reply, but if it's more than 50 lines of quoting, trigger a
"at least equal parts new message".

I'm sure Konstantin has nothing better to do...

                Linus

Re: KTODO automated TODO lists

[Geert Uytterhoeven]

Hi Linus,

On Mon, Oct 23, 2023 at 8:56 PM Linus Torvalds
<torvalds@linux-foundation.org> wrote:
> On Mon, 23 Oct 2023 at 08:49, Andrew Morton <akpm@linux-foundation.org> wrote:
> > Well here's a task: write a bot which follows the mailing lists and
> > sends people nastygrams if one of their emails is more than 95%(?)
> > quoted text.
>
> I think that might be better off as a spam filter rule.
>
> Don't make it some after-the-fact "trawl the lists". Just make it a
> bounce with a "you quoted too much". Same as the html avoidance.
>
> Make it ok to quote 15 lines of commit message for a "Reviewed-by:"
> kind of reply, but if it's more than 50 lines of quoting, trigger a
> "at least equal parts new message".

How to handle the (unfortunately fairly common) case of
reply-with-CC-of-forgotten-relevant-person_added?

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: KTODO automated TODO lists

[Linus Torvalds]

On Mon, 23 Oct 2023 at 09:00, Geert Uytterhoeven <geert@linux-m68k.org> wrote:
>
> How to handle the (unfortunately fairly common) case of
> reply-with-CC-of-forgotten-relevant-person_added?

One option might be to just do a proper non-quoting reply and have a
link to lore for context for the originally-forgotten person.

I've done that anyway several times, where I reply to something, and
bring in new people, and add a lore link for context.

I wouldn't have recommended that a few years ago, but lore has been so
reliable (and reliably fast - not usually the case for most email
archives) that I think it's a great way to fill people in.

But yes, it does end up being a "outside email" thing. I suspect we're
all good with that these days - the days of people working in text
terminals purely out of email are long gone anyway, methinks.

Anyway, I'm in no way trying to say this is the way forward, but I do
agree with Andrew that we've seen quite a bit of "long messages that
are mostly quoting", and it's not great.

              Linus

Re: KTODO automated TODO lists

[Steven Rostedt]

On Mon, 23 Oct 2023 08:55:56 -1000
Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Mon, 23 Oct 2023 at 08:49, Andrew Morton <akpm@linux-foundation.org> wrote:
> >
> > Well here's a task: write a bot which follows the mailing lists and
> > sends people nastygrams if one of their emails is more than 95%(?)
> > quoted text.  
> 
> I think that might be better off as a spam filter rule.
> 
> Don't make it some after-the-fact "trawl the lists". Just make it a
> bounce with a "you quoted too much". Same as the html avoidance.
> 
> Make it ok to quote 15 lines of commit message for a "Reviewed-by:"
> kind of reply, but if it's more than 50 lines of quoting, trigger a
> "at least equal parts new message".
> 
> I'm sure Konstantin has nothing better to do...
> 
>                 Linus

Paul,

Just in case you are wondering why one day one of your replies gets
rejected ;-)

-- Steve

Re: KTODO automated TODO lists

[Konstantin Ryabitsev]

On Mon, Oct 23, 2023 at 08:55:56AM -1000, Linus Torvalds wrote:
> > Well here's a task: write a bot which follows the mailing lists and
> > sends people nastygrams if one of their emails is more than 95%(?)
> > quoted text.
> 
> I think that might be better off as a spam filter rule.
> 
> Don't make it some after-the-fact "trawl the lists". Just make it a
> bounce with a "you quoted too much". Same as the html avoidance.

I know people aren't being very serious, but automating this away either
aggressively (reject as spam) or passive-agressively (whine at poster) will
run into rare but valid corner cases. For example, we have no way of
distinguishing between "this person quoted too much from previous message" and
"this person posted a large but relevant quote from docs or another
conversation," and so we will likely punish/annoy the innocent.

It's better to treat this as a mentoring opportunity and send an off-list
reply with "please trim your quotes" and maybe a link to
https://people.kernel.org/tglx/notes-about-netiquette

If it helps, I can add a mailing list etiquette page on subspace.kernel.org,
so it's easier to find.

-K

Re: KTODO automated TODO lists

[Paul E. McKenney]

On Mon, Oct 23, 2023 at 03:29:18PM -0400, Steven Rostedt wrote:
> On Mon, 23 Oct 2023 08:55:56 -1000
> Linus Torvalds <torvalds@linux-foundation.org> wrote:
> 
> > On Mon, 23 Oct 2023 at 08:49, Andrew Morton <akpm@linux-foundation.org> wrote:
> > >
> > > Well here's a task: write a bot which follows the mailing lists and
> > > sends people nastygrams if one of their emails is more than 95%(?)
> > > quoted text.  
> > 
> > I think that might be better off as a spam filter rule.
> > 
> > Don't make it some after-the-fact "trawl the lists". Just make it a
> > bounce with a "you quoted too much". Same as the html avoidance.
> > 
> > Make it ok to quote 15 lines of commit message for a "Reviewed-by:"
> > kind of reply, but if it's more than 50 lines of quoting, trigger a
> > "at least equal parts new message".
> > 
> > I'm sure Konstantin has nothing better to do...
> > 
> >                 Linus
> 
> Paul,
> 
> Just in case you are wondering why one day one of your replies gets
> rejected ;-)

You never know.  Those who would have otherwise received my replies
might be very happy with this outcome.  ;-)

							Thanx, Paul

Re: KTODO automated TODO lists

[Tony Luck]

On Mon, Oct 23, 2023 at 02:31:11PM -0700, Paul E. McKenney wrote:
> On Mon, Oct 23, 2023 at 03:29:18PM -0400, Steven Rostedt wrote:
> > On Mon, 23 Oct 2023 08:55:56 -1000
> > Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > 
> > > On Mon, 23 Oct 2023 at 08:49, Andrew Morton <akpm@linux-foundation.org> wrote:
> > > >
> > > > Well here's a task: write a bot which follows the mailing lists and
> > > > sends people nastygrams if one of their emails is more than 95%(?)
> > > > quoted text.  
> > > 
> > > I think that might be better off as a spam filter rule.
> > > 
> > > Don't make it some after-the-fact "trawl the lists". Just make it a
> > > bounce with a "you quoted too much". Same as the html avoidance.
> > > 
> > > Make it ok to quote 15 lines of commit message for a "Reviewed-by:"
> > > kind of reply, but if it's more than 50 lines of quoting, trigger a
> > > "at least equal parts new message".
> > > 
> > > I'm sure Konstantin has nothing better to do...
> > > 
> > >                 Linus
> > 
> > Paul,
> > 
> > Just in case you are wondering why one day one of your replies gets
> > rejected ;-)
> 
> You never know.  Those who would have otherwise received my replies
> might be very happy with this outcome.  ;-)
> 
> 							Thanx, Paul

Hmm.

Thirty-two lines of quoted message.

Only two lines of response.

[Not including signature]

You are skating close to the edge of a 95% quote rule filter unless
it counted the signature.

But
this
might
also
cause
people
to
go
to
silly
lengths
to
avoid
having
their
message
cancelled!

-Tony

Re: KTODO automated TODO lists

[NeilBrown]

On Tue, 24 Oct 2023, Andrew Morton wrote:
> On Thu, 19 Oct 2023 07:11:36 +0300 Dan Carpenter <dan.carpenter@linaro.org> wrote:
> 
> > Yesterday someone on my lists just sent an email looking for kernel
> > tasks. 
> 
> Well here's a task: write a bot which follows the mailing lists and
> sends people nastygrams if one of their emails is more than 95%(?)
> quoted text.

Doesn't your email reader automatically hide most of a large quote?
Mine does :-)

NeilBrown


> 
> It's happening significantly more lately.  Possibly because the gmail
> client helpfully hides quoted text.
> 
> Probably not a great way of becoming popular.
> 
> 

Re: KTODO automated TODO lists

[Paul E. McKenney]

On Mon, Oct 23, 2023 at 02:44:15PM -0700, Tony Luck wrote:
> On Mon, Oct 23, 2023 at 02:31:11PM -0700, Paul E. McKenney wrote:
> > On Mon, Oct 23, 2023 at 03:29:18PM -0400, Steven Rostedt wrote:
> > > On Mon, 23 Oct 2023 08:55:56 -1000
> > > Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > > 
> > > > On Mon, 23 Oct 2023 at 08:49, Andrew Morton <akpm@linux-foundation.org> wrote:
> > > > >
> > > > > Well here's a task: write a bot which follows the mailing lists and
> > > > > sends people nastygrams if one of their emails is more than 95%(?)
> > > > > quoted text.  
> > > > 
> > > > I think that might be better off as a spam filter rule.
> > > > 
> > > > Don't make it some after-the-fact "trawl the lists". Just make it a
> > > > bounce with a "you quoted too much". Same as the html avoidance.
> > > > 
> > > > Make it ok to quote 15 lines of commit message for a "Reviewed-by:"
> > > > kind of reply, but if it's more than 50 lines of quoting, trigger a
> > > > "at least equal parts new message".
> > > > 
> > > > I'm sure Konstantin has nothing better to do...
> > > > 
> > > >                 Linus
> > > 
> > > Paul,
> > > 
> > > Just in case you are wondering why one day one of your replies gets
> > > rejected ;-)
> > 
> > You never know.  Those who would have otherwise received my replies
> > might be very happy with this outcome.  ;-)
> > 
> > 							Thanx, Paul
> 
> Hmm.
> 
> Thirty-two lines of quoted message.
> 
> Only two lines of response.
> 
> [Not including signature]
> 
> You are skating close to the edge of a 95% quote rule filter unless
> it counted the signature.
> 
> But
> this
> might
> also
> cause
> people
> to
> go
> to
> silly
> lengths
> to
> avoid
> having
> their
> message
> cancelled!

;-)
;-)
;-)
;-)
;-)
;-)
;-)
;-)
;-)
;-)
;-)
;-)
;-)
;-)
;-)
;-)
;-)
;-)
;-)
;-)
;-)
;-)
;-)
;-)

							Thanx, Paul

Re: KTODO automated TODO lists

[Gustavo A. R. Silva]

On 10/18/23 22:11, Dan Carpenter wrote:
> Yesterday someone on my lists just sent an email looking for kernel
> tasks.  This was a university student in a kernel programming class.
> We also have kernel-janitors and outreachy and those people are always
> asking for small tasks.

We have tons of issues waiting to be audited and fixed here:

https://scan.coverity.com/projects/linux-next-weekly-scan

You will never run out of fun. :) People just need to sign up.

That's really a great way to learn and gain experience across the whole
kernel tree.

--
Gustavo

Re: KTODO automated TODO lists

[Gustavo A. R. Silva]

On 10/23/23 17:38, Gustavo A. R. Silva wrote:
> 
> 
> On 10/18/23 22:11, Dan Carpenter wrote:
>> Yesterday someone on my lists just sent an email looking for kernel
>> tasks.  This was a university student in a kernel programming class.
>> We also have kernel-janitors and outreachy and those people are always
>> asking for small tasks.
> 
> We have tons of issues waiting to be audited and fixed here:
> 
> https://scan.coverity.com/projects/linux-next-weekly-scan
> 
> You will never run out of fun. :) People just need to sign up.
I see people already signing up. :)

BTW, here are some resources that people could find valuable:

https://embeddedor.com/slides/2017/kr/kr2017.pdf
https://embeddedor.com/slides/2018/kr/kr2018.pdf
https://embeddedor.com/slides/2019/kr/kr2019.pdf

Those are the slides from some presentations about fixing Coverity (and other)
issues that I gave at Kernel Recipes some years ago. The talks are also on
YouTube if people want to take a look.

--
Gustavo

Re: KTODO automated TODO lists

[Joe Perches]

On Mon, 2023-10-23 at 17:38 -0600, Gustavo A. R. Silva wrote:
> We have tons of issues waiting to be audited and fixed here:
> 
> https://scan.coverity.com/projects/linux-next-weekly-scan
> 
> You will never run out of fun. :) People just need to sign up.

IMO: People should not need to sign up.

Re: KTODO automated TODO lists

[Andrew Morton]

On Mon, 23 Oct 2023 15:41:48 -0400 Konstantin Ryabitsev <mricon@kernel.org> wrote:

> On Mon, Oct 23, 2023 at 08:55:56AM -1000, Linus Torvalds wrote:
> > > Well here's a task: write a bot which follows the mailing lists and
> > > sends people nastygrams if one of their emails is more than 95%(?)
> > > quoted text.
> > 
> > I think that might be better off as a spam filter rule.
> > 
> > Don't make it some after-the-fact "trawl the lists". Just make it a
> > bounce with a "you quoted too much". Same as the html avoidance.
> 
> I know people aren't being very serious, but automating this away either
> aggressively (reject as spam) or passive-agressively (whine at poster) will
> run into rare but valid corner cases. For example, we have no way of
> distinguishing between "this person quoted too much from previous message" and
> "this person posted a large but relevant quote from docs or another
> conversation," and so we will likely punish/annoy the innocent.

Rejecting a legtimate email would be bad.

So we choose "whine at poster".  If it's a false positive then they'll
somehow survive the experience.   And, most importantly, the mail
will get through.

> It's better to treat this as a mentoring opportunity and send an off-list
> reply with "please trim your quotes" and maybe a link to
> https://people.kernel.org/tglx/notes-about-netiquette

I'm tiring of sending off-list emails.  This task could be automated. 
Which is what I'm suggesting!

> If it helps, I can add a mailing list etiquette page on subspace.kernel.org,
> so it's easier to find.

Great, link to that in the whiney emails.

Re: KTODO automated TODO lists

[Geert Uytterhoeven]

Hi Neil,

On Mon, Oct 23, 2023 at 11:46 PM NeilBrown <neilb@suse.de> wrote:
> On Tue, 24 Oct 2023, Andrew Morton wrote:
> > On Thu, 19 Oct 2023 07:11:36 +0300 Dan Carpenter <dan.carpenter@linaro.org> wrote:
> > > Yesterday someone on my lists just sent an email looking for kernel
> > > tasks.
> >
> > Well here's a task: write a bot which follows the mailing lists and
> > sends people nastygrams if one of their emails is more than 95%(?)
> > quoted text.
>
> Doesn't your email reader automatically hide most of a large quote?
> Mine does :-)

That's part of the problem: many people don't see anymore if the
previous email author removed irrelevant parts or not.  Until they
want to reply...

> > It's happening significantly more lately.  Possibly because the gmail
> > client helpfully hides quoted text.

When replying, the Gmail web interface (or Chrome?) is also very
slow when selecting very long irrelevant parts for deletion.  And it's
hard to predict when "Show original" and "b4 mbox && alpine -f"
would be faster...

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: KTODO automated TODO lists

[Laurent Pinchart]

Hi Geert,

On Tue, Oct 24, 2023 at 09:19:26AM +0200, Geert Uytterhoeven wrote:
> On Mon, Oct 23, 2023 at 11:46 PM NeilBrown wrote:
> > On Tue, 24 Oct 2023, Andrew Morton wrote:
> > > On Thu, 19 Oct 2023 07:11:36 +0300 Dan Carpenter wrote:
> > > > Yesterday someone on my lists just sent an email looking for kernel
> > > > tasks.
> > >
> > > Well here's a task: write a bot which follows the mailing lists and
> > > sends people nastygrams if one of their emails is more than 95%(?)
> > > quoted text.
> >
> > Doesn't your email reader automatically hide most of a large quote?
> > Mine does :-)
> 
> That's part of the problem: many people don't see anymore if the
> previous email author removed irrelevant parts or not.  Until they
> want to reply...
> 
> > > It's happening significantly more lately.  Possibly because the gmail
> > > client helpfully hides quoted text.
> 
> When replying, the Gmail web interface (or Chrome?) is also very
> slow when selecting very long irrelevant parts for deletion.  And it's
> hard to predict when "Show original" and "b4 mbox && alpine -f"
> would be faster...

Get a better e-mail client ? ;-) At least with e-mail you have a choice
between different clients.

I've refrained from replying to this thread so far, as it seemed to be a
caricature of a bikeshedding discussion, but for what it's worth, I
often find myself in the opposite situation when I'm annoyed that
someone trimmed too much of the discussion in their replies.

Yes, replying to a 3000-lines patches with a full quote ana d a
Reviewed-by tag at the very bottom, without any other comment, is
annoying. On the other hand, trimming everything but the few lines to
which you reply means that it gets much more annoying to jump in the
discussion in the middle of a mail thread. There's a difference between
trimming unrelated parts, and removing related content that happens not
to be the direct subject of a particular reply.

-- 
Regards,

Laurent Pinchart

Re: KTODO automated TODO lists

[Jani Nikula]

On Tue, 24 Oct 2023, Laurent Pinchart <laurent.pinchart@ideasonboard.com> wrote:
> Get a better e-mail client ? ;-) At least with e-mail you have a choice
> between different clients.

Yup. What I see is excessive quotes collapsed, replaced with something
like this:

[ 18 more citation lines. Click/Enter to show. ]

All the actual replies stand out, regardless of the length of
quoting. Now it's just the Outlook style "quoting" without >'s that bugs
me...


BR,
Jani.


-- 
Jani Nikula, Intel

Re: KTODO automated TODO lists

[Laurent Pinchart]

On Tue, Oct 24, 2023 at 11:42:26AM +0300, Jani Nikula wrote:
> On Tue, 24 Oct 2023, Laurent Pinchart wrote:
> > Get a better e-mail client ? ;-) At least with e-mail you have a choice
> > between different clients.
> 
> Yup. What I see is excessive quotes collapsed, replaced with something
> like this:
> 
> [ 18 more citation lines. Click/Enter to show. ]
> 
> All the actual replies stand out, regardless of the length of
> quoting. Now it's just the Outlook style "quoting" without >'s that bugs
> me...

On the list of tasks I will never get to is the development of a mail
filter that turns outlook-style "quoting" into real quoting when
replying to e-mails (in mutt in my case). Does anyone know of anything
like that having been developed already ?

-- 
Regards,

Laurent Pinchart

Re: KTODO automated TODO lists

[Steven Rostedt]

On Tue, 24 Oct 2023 10:25:06 +0300
Laurent Pinchart <laurent.pinchart@ideasonboard.com> wrote:

> I've refrained from replying to this thread so far, as it seemed to be a
> caricature of a bikeshedding discussion, but for what it's worth, I
> often find myself in the opposite situation when I'm annoyed that
> someone trimmed too much of the discussion in their replies.

After hitting "page down" 3 or 4 times and seeing only quoted text, I then
stop and just ignore the email. Yes, there's been emails I purposely
ignored because of this that had asked me to respond near the end. Oh well.
Then they ask, "why didn't you respond?" pointing out the email I was to
respond to. And I would reply, "I never saw the request because of too much
quoted text".

> 
> Yes, replying to a 3000-lines patches with a full quote ana d a
> Reviewed-by tag at the very bottom, without any other comment, is
> annoying. On the other hand, trimming everything but the few lines to
> which you reply means that it gets much more annoying to jump in the
> discussion in the middle of a mail thread. There's a difference between
> trimming unrelated parts, and removing related content that happens not
> to be the direct subject of a particular reply.

I just replied to an email yesterday that cut too much off, and I had to
make a note about that, and put things back in.

What's worse, is if you are having a technical debate with someone, and
they trim out everything that might go against their argument, but leave
anything that supports their argument. I've seen that happen quite a bit.

I should write a book called "The art of trimming". ;-)

-- Steve

Re: KTODO automated TODO lists

[Laurent Pinchart]

On Tue, Oct 24, 2023 at 08:36:47AM -0400, Steven Rostedt wrote:
> On Tue, 24 Oct 2023 10:25:06 +0300 Laurent Pinchart wrote:
> 
> > I've refrained from replying to this thread so far, as it seemed to be a
> > caricature of a bikeshedding discussion, but for what it's worth, I
> > often find myself in the opposite situation when I'm annoyed that
> > someone trimmed too much of the discussion in their replies.
> 
> After hitting "page down" 3 or 4 times and seeing only quoted text, I then
> stop and just ignore the email. Yes, there's been emails I purposely
> ignored because of this that had asked me to respond near the end. Oh well.
> Then they ask, "why didn't you respond?" pointing out the email I was to
> respond to. And I would reply, "I never saw the request because of too much
> quoted text".
> 
> > Yes, replying to a 3000-lines patches with a full quote ana d a
> > Reviewed-by tag at the very bottom, without any other comment, is
> > annoying. On the other hand, trimming everything but the few lines to
> > which you reply means that it gets much more annoying to jump in the
> > discussion in the middle of a mail thread. There's a difference between
> > trimming unrelated parts, and removing related content that happens not
> > to be the direct subject of a particular reply.
> 
> I just replied to an email yesterday that cut too much off, and I had to
> make a note about that, and put things back in.
> 
> What's worse, is if you are having a technical debate with someone, and
> they trim out everything that might go against their argument, but leave
> anything that supports their argument. I've seen that happen quite a bit.
> 
> I should write a book called "The art of trimming". ;-)

Maybe a good path forward would be to start by flagging extreme cases
only, without being too pedantic ? That assumes we can agree what an
extreme case is.

One thing I found helpful in replies is to add tags just after the
commit message (where the tag will appear when it gets applied), or
after the last comment if I need to comment on something specific. The
recipient will know that they don't need to scroll down after the tag.

-- 
Regards,

Laurent Pinchart

Re: KTODO automated TODO lists

[Konstantin Ryabitsev]

On Mon, Oct 23, 2023 at 03:41:48PM -0400, Konstantin Ryabitsev wrote:
> If it helps, I can add a mailing list etiquette page on subspace.kernel.org,
> so it's easier to find.

FWIW, such page now exists:
https://subspace.kernel.org/etiquette.html

-K

Re: KTODO automated TODO lists

[Krzysztof Kozlowski]

On 23/10/2023 20:49, Andrew Morton wrote:
> On Thu, 19 Oct 2023 07:11:36 +0300 Dan Carpenter <dan.carpenter@linaro.org> wrote:
> 
>> Yesterday someone on my lists just sent an email looking for kernel
>> tasks. 
> 
> Well here's a task: write a bot which follows the mailing lists and
> sends people nastygrams if one of their emails is more than 95%(?)
> quoted text.
> 
> It's happening significantly more lately.  Possibly because the gmail
> client helpfully hides quoted text.

I would also point to reviewers and maintainers who give a Rb/Ack tag:
1. somewhere at the top, without any footer like Best regards, and then
quote entire patch, so I don't know shall I look for more comments after
Rb/Ack?

2. quote entire email and then add Rb/Ack, so I need to figure out
whether there was something between the hundreds of lines of text or not.

Best regards,
Krzysztof

Re: KTODO automated TODO lists

[NeilBrown]

On Wed, 25 Oct 2023, Krzysztof Kozlowski wrote:
> On 23/10/2023 20:49, Andrew Morton wrote:
> > On Thu, 19 Oct 2023 07:11:36 +0300 Dan Carpenter <dan.carpenter@linaro.org> wrote:
> > 
> >> Yesterday someone on my lists just sent an email looking for kernel
> >> tasks. 
> > 
> > Well here's a task: write a bot which follows the mailing lists and
> > sends people nastygrams if one of their emails is more than 95%(?)
> > quoted text.
> > 
> > It's happening significantly more lately.  Possibly because the gmail
> > client helpfully hides quoted text.
> 
> I would also point to reviewers and maintainers who give a Rb/Ack tag:
> 1. somewhere at the top, without any footer like Best regards, and then
> quote entire patch, so I don't know shall I look for more comments after
> Rb/Ack?
> 
> 2. quote entire email and then add Rb/Ack, so I need to figure out
> whether there was something between the hundreds of lines of text or not.

Here we all are, brilliantly talented computer programmers who spend
our days making amazing fast digital devices do amazingly clever and
subtle things, inventing time-saving tools and processing vast amounts
of data without blinking, but for some reason we think the task of
skipping over a few thousand lines that all start with '> " is too hard
for us and that we should, in stead, complain to some other human to
convince them to make our life easier for us.

Does anyone else see the irony?

NeilBrown

Re: KTODO automated TODO lists

[Steven Rostedt]

On Wed, 25 Oct 2023 08:29:14 +1100
"NeilBrown" <neilb@suse.de> wrote:

> Here we all are, brilliantly talented computer programmers who spend
> our days making amazing fast digital devices do amazingly clever and
> subtle things, inventing time-saving tools and processing vast amounts
> of data without blinking, but for some reason we think the task of
> skipping over a few thousand lines that all start with '> " is too hard
> for us and that we should, in stead, complain to some other human to
> convince them to make our life easier for us.
> 
> Does anyone else see the irony?

Did you also know that real-time developers are the most unpredictable?

-- Steve

Re: KTODO automated TODO lists

[Paul E. McKenney]

On Tue, Oct 24, 2023 at 06:05:17PM -0400, Steven Rostedt wrote:
> On Wed, 25 Oct 2023 08:29:14 +1100
> "NeilBrown" <neilb@suse.de> wrote:
> 
> > Here we all are, brilliantly talented computer programmers who spend
> > our days making amazing fast digital devices do amazingly clever and
> > subtle things, inventing time-saving tools and processing vast amounts
> > of data without blinking, but for some reason we think the task of
> > skipping over a few thousand lines that all start with '> " is too hard
> > for us and that we should, in stead, complain to some other human to
> > convince them to make our life easier for us.
> > 
> > Does anyone else see the irony?
> 
> Did you also know that real-time developers are the most unpredictable?

Are safety-critical programmers the most easy-going?

							Thanx, Paul

Re: KTODO automated TODO lists

[Geert Uytterhoeven]

Hi Neil,

On Tue, Oct 24, 2023 at 11:29 PM NeilBrown <neilb@suse.de> wrote:
> On Wed, 25 Oct 2023, Krzysztof Kozlowski wrote:
> > On 23/10/2023 20:49, Andrew Morton wrote:
> > > On Thu, 19 Oct 2023 07:11:36 +0300 Dan Carpenter <dan.carpenter@linaro.org> wrote:
> > >> Yesterday someone on my lists just sent an email looking for kernel
> > >> tasks.
> > >
> > > Well here's a task: write a bot which follows the mailing lists and
> > > sends people nastygrams if one of their emails is more than 95%(?)
> > > quoted text.
> > >
> > > It's happening significantly more lately.  Possibly because the gmail
> > > client helpfully hides quoted text.
> >
> > I would also point to reviewers and maintainers who give a Rb/Ack tag:
> > 1. somewhere at the top, without any footer like Best regards, and then
> > quote entire patch, so I don't know shall I look for more comments after
> > Rb/Ack?
> >
> > 2. quote entire email and then add Rb/Ack, so I need to figure out
> > whether there was something between the hundreds of lines of text or not.
>
> Here we all are, brilliantly talented computer programmers who spend
> our days making amazing fast digital devices do amazingly clever and
> subtle things, inventing time-saving tools and processing vast amounts
> of data without blinking, but for some reason we think the task of
> skipping over a few thousand lines that all start with '> " is too hard
> for us and that we should, in stead, complain to some other human to
> convince them to make our life easier for us.
>
> Does anyone else see the irony?

Please compare the numbers:
  1. 1 sender removes irrelevant parts,
  2. N receivers skip irrelevant parts.

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: KTODO automated TODO lists

[Krzysztof Kozlowski]

On 24/10/2023 23:29, NeilBrown wrote:
> On Wed, 25 Oct 2023, Krzysztof Kozlowski wrote:
>> On 23/10/2023 20:49, Andrew Morton wrote:
>>> On Thu, 19 Oct 2023 07:11:36 +0300 Dan Carpenter <dan.carpenter@linaro.org> wrote:
>>>
>>>> Yesterday someone on my lists just sent an email looking for kernel
>>>> tasks. 
>>>
>>> Well here's a task: write a bot which follows the mailing lists and
>>> sends people nastygrams if one of their emails is more than 95%(?)
>>> quoted text.
>>>
>>> It's happening significantly more lately.  Possibly because the gmail
>>> client helpfully hides quoted text.
>>
>> I would also point to reviewers and maintainers who give a Rb/Ack tag:
>> 1. somewhere at the top, without any footer like Best regards, and then
>> quote entire patch, so I don't know shall I look for more comments after
>> Rb/Ack?
>>
>> 2. quote entire email and then add Rb/Ack, so I need to figure out
>> whether there was something between the hundreds of lines of text or not.
> 
> Here we all are, brilliantly talented computer programmers who spend
> our days making amazing fast digital devices do amazingly clever and
> subtle things, inventing time-saving tools and processing vast amounts
> of data without blinking, but for some reason we think the task of
> skipping over a few thousand lines that all start with '> " is too hard
> for us and that we should, in stead, complain to some other human to
> convince them to make our life easier for us.
> 
> Does anyone else see the irony?

Neither my email client is able to skip the quoted text, nor
lore.kernel.org or Patchwork. Therefore in both cases - reading the
email with my current toolset or checking some older threads on
lore/Patchwork - I would have to do it manually. Quite a waste of time.

Best regards,
Krzysztof

Re: KTODO automated TODO lists

[James Bottomley]

On Wed, 2023-10-25 at 08:29 +1100, NeilBrown wrote:
> On Wed, 25 Oct 2023, Krzysztof Kozlowski wrote:
> > On 23/10/2023 20:49, Andrew Morton wrote:
> > > On Thu, 19 Oct 2023 07:11:36 +0300 Dan Carpenter
> > > <dan.carpenter@linaro.org> wrote:
> > > 
> > > > Yesterday someone on my lists just sent an email looking for
> > > > kernel
> > > > tasks. 
> > > 
> > > Well here's a task: write a bot which follows the mailing lists
> > > and
> > > sends people nastygrams if one of their emails is more than
> > > 95%(?)
> > > quoted text.
> > > 
> > > It's happening significantly more lately.  Possibly because the
> > > gmail
> > > client helpfully hides quoted text.
> > 
> > I would also point to reviewers and maintainers who give a Rb/Ack
> > tag:
> > 1. somewhere at the top, without any footer like Best regards, and
> > then
> > quote entire patch, so I don't know shall I look for more comments
> > after
> > Rb/Ack?
> > 
> > 2. quote entire email and then add Rb/Ack, so I need to figure out
> > whether there was something between the hundreds of lines of text
> > or not.
> 
> Here we all are, brilliantly talented computer programmers who spend
> our days making amazing fast digital devices do amazingly clever and
> subtle things, inventing time-saving tools and processing vast
> amounts of data without blinking, but for some reason we think the
> task of skipping over a few thousand lines that all start with '> "
> is too hard for us and that we should, in stead, complain to some
> other human to convince them to make our life easier for us.
> 
> Does anyone else see the irony?

So if I'm a brilliantly talented driver, it's OK for other people to
drive on the wrong side of the road because I should be able to avoid
them?

The point being there are some situations where observing global
etiquette is way more helpful than an individual solution, however
talented the individual.

James

Re: KTODO automated TODO lists

[Jani Nikula]

On Wed, 25 Oct 2023, James Bottomley <James.Bottomley@HansenPartnership.com> wrote:
> On Wed, 2023-10-25 at 08:29 +1100, NeilBrown wrote:
>> Here we all are, brilliantly talented computer programmers who spend
>> our days making amazing fast digital devices do amazingly clever and
>> subtle things, inventing time-saving tools and processing vast
>> amounts of data without blinking, but for some reason we think the
>> task of skipping over a few thousand lines that all start with '> "
>> is too hard for us and that we should, in stead, complain to some
>> other human to convince them to make our life easier for us.
>> 
>> Does anyone else see the irony?
>
> So if I'm a brilliantly talented driver, it's OK for other people to
> drive on the wrong side of the road because I should be able to avoid
> them?

Nah, we're all brilliant car manufacturers that could have our cars deal
with the situation. ;)

The notmuch emacs interface has collapsed citations since at least
2010. I think Neil's point is, if we're all using open source MUAs, why
don't we scratch that particular itch and move on, instead of getting
frustrated by it year after year?


BR,
Jani.


-- 
Jani Nikula, Intel

Re: KTODO automated TODO lists

[James Bottomley]

On Wed, 2023-10-25 at 19:40 +0300, Jani Nikula wrote:
> On Wed, 25 Oct 2023, James Bottomley
> <James.Bottomley@HansenPartnership.com> wrote:
> > On Wed, 2023-10-25 at 08:29 +1100, NeilBrown wrote:
> > > Here we all are, brilliantly talented computer programmers who
> > > spend our days making amazing fast digital devices do amazingly
> > > clever and subtle things, inventing time-saving tools and
> > > processing vast amounts of data without blinking, but for some
> > > reason we think the task of skipping over a few thousand lines
> > > that all start with '>"s too hard for us and that we should, in
> > > stead, complain to some other human to convince them to make our
> > > life easier for us.
> > > 
> > > Does anyone else see the irony?
> > 
> > So if I'm a brilliantly talented driver, it's OK for other people
> > to drive on the wrong side of the road because I should be able to
> > avoid them?
> 
> Nah, we're all brilliant car manufacturers that could have our cars
> deal with the situation. ;)
> 
> The notmuch emacs interface has collapsed citations since at least
> 2010. I think Neil's point is, if we're all using open source MUAs,
> why don't we scratch that particular itch and move on, instead of
> getting frustrated by it year after year?

Because some MUAs don't have it.  Others are a bit aggressive, meaning
you have to turn it off anyway if you want to see what's in more than a
couple of lines of a quote (it only takes me a couple of emails to get
incredibly annoyed with the way gmail does it, for instance, since it
never seems to leave enough useful context).

James

Re: KTODO automated TODO lists

[Steven Rostedt]

On Wed, 25 Oct 2023 14:07:02 -0400
James Bottomley <James.Bottomley@HansenPartnership.com> wrote:

> Because some MUAs don't have it.  Others are a bit aggressive, meaning
> you have to turn it off anyway if you want to see what's in more than a
> couple of lines of a quote (it only takes me a couple of emails to get
> incredibly annoyed with the way gmail does it, for instance, since it
> never seems to leave enough useful context).

I think this is the key issue. We only want the context of an email that is
being responded to present, and the rest trimmed. Automated trimming or
collapsing doesn't do this well.

Maybe we can make AI do this for us.. hmm

-- Steve

Re: KTODO automated TODO lists

[Alexey Dobriyan]

On Wed, Oct 25, 2023 at 07:45:55AM -0400, James Bottomley wrote:
> On Wed, 2023-10-25 at 08:29 +1100, NeilBrown wrote:
> > On Wed, 25 Oct 2023, Krzysztof Kozlowski wrote:
> > > On 23/10/2023 20:49, Andrew Morton wrote:
> > > > On Thu, 19 Oct 2023 07:11:36 +0300 Dan Carpenter
> > > > <dan.carpenter@linaro.org> wrote:
> > > > 
> > > > > Yesterday someone on my lists just sent an email looking for
> > > > > kernel
> > > > > tasks. 
> > > > 
> > > > Well here's a task: write a bot which follows the mailing lists
> > > > and
> > > > sends people nastygrams if one of their emails is more than
> > > > 95%(?)
> > > > quoted text.
> > > > 
> > > > It's happening significantly more lately.  Possibly because the
> > > > gmail
> > > > client helpfully hides quoted text.
> > > 
> > > I would also point to reviewers and maintainers who give a Rb/Ack
> > > tag:
> > > 1. somewhere at the top, without any footer like Best regards, and
> > > then
> > > quote entire patch, so I don't know shall I look for more comments
> > > after
> > > Rb/Ack?
> > > 
> > > 2. quote entire email and then add Rb/Ack, so I need to figure out
> > > whether there was something between the hundreds of lines of text
> > > or not.
> > 
> > Here we all are, brilliantly talented computer programmers who spend
> > our days making amazing fast digital devices do amazingly clever and
> > subtle things, inventing time-saving tools and processing vast
> > amounts of data without blinking, but for some reason we think the
> > task of skipping over a few thousand lines that all start with '> "
> > is too hard for us and that we should, in stead, complain to some
> > other human to convince them to make our life easier for us.
> > 
> > Does anyone else see the irony?
> 
> So if I'm a brilliantly talented driver, it's OK for other people to
> drive on the wrong side of the road because I should be able to avoid
> them?
> 
> The point being there are some situations where observing global
> etiquette is way more helpful than an individual solution, however
> talented the individual.

It's MUAs failure. Microsoft "solved" this problem by forcing top
posting onto everyone, but there is no reason Outlook couldn't scroll
to the first reply in the middle of email. I use mutt, it doesn't
scroll to the first reply either.

Re: KTODO automated TODO lists

[Steven Rostedt]

On Wed, 25 Oct 2023 21:55:16 +0300
Alexey Dobriyan <adobriyan@gmail.com> wrote:

> It's MUAs failure. Microsoft "solved" this problem by forcing top
> posting onto everyone, but there is no reason Outlook couldn't scroll
> to the first reply in the middle of email. I use mutt, it doesn't
> scroll to the first reply either.

Outlook is the worse. There may be no reason it can't scroll to the first
reply, but in many cases it just hides it! When I worked at VMware, I would
always inline my replies (and I was forced to use outlook). If I didn't
remove the:

  On Wed, 25 Oct 2023 21:55:16 +0300
  Jane Doe <jane.doe@vmware.com> wrote:

Line at the top, it would show a blank message. I had to manually start
removing that when I received several responses telling me "You sent me a
blank email"!

-- Steve

Re: KTODO automated TODO lists

[Linus Torvalds]

On Wed, 25 Oct 2023 at 08:10, Steven Rostedt <rostedt@goodmis.org> wrote:
>
> I think this is the key issue. We only want the context of an email that is
> being responded to present, and the rest trimmed. Automated trimming or
> collapsing doesn't do this well.

It's not just about MUA's that hide things. The MUA _I_ use hides
excessive quoting, but then when I look at the results on lore I often
get completely unreadable results because somebody quoted several
hundred lines of patch or whatever. And then I scrolling through the
thread is suddenly a major PITA.

So I do think that the whole excessive quoting on the lists should
just be a hard no, the same way html is.

                 Linus

Re: KTODO automated TODO lists

[Alexandre Belloni]

On 25/10/2023 21:55:16+0300, Alexey Dobriyan wrote:
> It's MUAs failure. Microsoft "solved" this problem by forcing top
> posting onto everyone, but there is no reason Outlook couldn't scroll
> to the first reply in the middle of email. I use mutt, it doesn't
> scroll to the first reply either.
> 

Just press 'S'?

<skip-quoted> (default: S)
This function will go to the next line of non-quoted text which comes after a line of quoted text in the internal pager.

-- 
Alexandre Belloni, co-owner and COO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com

Re: KTODO automated TODO lists

[NeilBrown]

On Wed, 25 Oct 2023, Geert Uytterhoeven wrote:
> Hi Neil,
> 
> On Tue, Oct 24, 2023 at 11:29 PM NeilBrown <neilb@suse.de> wrote:
> > On Wed, 25 Oct 2023, Krzysztof Kozlowski wrote:
> > > On 23/10/2023 20:49, Andrew Morton wrote:
> > > > On Thu, 19 Oct 2023 07:11:36 +0300 Dan Carpenter <dan.carpenter@linaro.org> wrote:
> > > >> Yesterday someone on my lists just sent an email looking for kernel
> > > >> tasks.
> > > >
> > > > Well here's a task: write a bot which follows the mailing lists and
> > > > sends people nastygrams if one of their emails is more than 95%(?)
> > > > quoted text.
> > > >
> > > > It's happening significantly more lately.  Possibly because the gmail
> > > > client helpfully hides quoted text.
> > >
> > > I would also point to reviewers and maintainers who give a Rb/Ack tag:
> > > 1. somewhere at the top, without any footer like Best regards, and then
> > > quote entire patch, so I don't know shall I look for more comments after
> > > Rb/Ack?
> > >
> > > 2. quote entire email and then add Rb/Ack, so I need to figure out
> > > whether there was something between the hundreds of lines of text or not.
> >
> > Here we all are, brilliantly talented computer programmers who spend
> > our days making amazing fast digital devices do amazingly clever and
> > subtle things, inventing time-saving tools and processing vast amounts
> > of data without blinking, but for some reason we think the task of
> > skipping over a few thousand lines that all start with '> " is too hard
> > for us and that we should, in stead, complain to some other human to
> > convince them to make our life easier for us.
> >
> > Does anyone else see the irony?
> 
> Please compare the numbers:
>   1. 1 sender removes irrelevant parts,
>   2. N receivers skip irrelevant parts.

That is one way to look at the numbers.
Another is:

  12 - fix about a dozen MUAs to summaries quotes properly
  12000 - fix an unknownable number of people to quote just exactly the
          amount that their particular audience is going to want

and when it comes to fixing-code versus fixing-people, I know which this
community is better at.

I guess there is also the option

   1 - fix vger.kernel.org to reject postings from people who don't
        think and quote like "us", because we already have too many
        contributor and want to block the heretics

This is really just a form of the "platform problem" which lwn.net has
occasionally written about.  The "problem" is that we treat the platform
(library code or other infrastructure) as fixed and develop ugly hacks
in our own code to work around some shortcoming, rather the going into
the platform and fixing it once for everyone there.

NeilBrown

Re: KTODO automated TODO lists

[NeilBrown]

On Thu, 26 Oct 2023, Steven Rostedt wrote:
> On Wed, 25 Oct 2023 14:07:02 -0400
> James Bottomley <James.Bottomley@HansenPartnership.com> wrote:
> 
> > Because some MUAs don't have it.  Others are a bit aggressive, meaning
> > you have to turn it off anyway if you want to see what's in more than a
> > couple of lines of a quote (it only takes me a couple of emails to get
> > incredibly annoyed with the way gmail does it, for instance, since it
> > never seems to leave enough useful context).
> 
> I think this is the key issue. We only want the context of an email that is
> being responded to present, and the rest trimmed. Automated trimming or
> collapsing doesn't do this well.

Who exactly is "We"?  I don't think your description matches what I
want.

NeilBrown

> 
> Maybe we can make AI do this for us.. hmm
> 
> -- Steve
> 

Re: KTODO automated TODO lists

[NeilBrown]

On Thu, 26 Oct 2023, Linus Torvalds wrote:
> On Wed, 25 Oct 2023 at 08:10, Steven Rostedt <rostedt@goodmis.org> wrote:
> >
> > I think this is the key issue. We only want the context of an email that is
> > being responded to present, and the rest trimmed. Automated trimming or
> > collapsing doesn't do this well.
> 
> It's not just about MUA's that hide things. The MUA _I_ use hides
> excessive quoting, but then when I look at the results on lore I often
> get completely unreadable results because somebody quoted several
> hundred lines of patch or whatever. And then I scrolling through the
> thread is suddenly a major PITA.

So you don't think that lore is an MUA?
If lore didn't summaries the headers, that would be a bug.
But that fact that it doesn't summarise the quotes isn't?

NeilBrown


> 
> So I do think that the whole excessive quoting on the lists should
> just be a hard no, the same way html is.
> 
>                  Linus
> 

Re: KTODO automated TODO lists

[NeilBrown]

On Wed, 25 Oct 2023, James Bottomley wrote:
> On Wed, 2023-10-25 at 08:29 +1100, NeilBrown wrote:
> > On Wed, 25 Oct 2023, Krzysztof Kozlowski wrote:
> > > On 23/10/2023 20:49, Andrew Morton wrote:
> > > > On Thu, 19 Oct 2023 07:11:36 +0300 Dan Carpenter
> > > > <dan.carpenter@linaro.org> wrote:
> > > > 
> > > > > Yesterday someone on my lists just sent an email looking for
> > > > > kernel
> > > > > tasks. 
> > > > 
> > > > Well here's a task: write a bot which follows the mailing lists
> > > > and
> > > > sends people nastygrams if one of their emails is more than
> > > > 95%(?)
> > > > quoted text.
> > > > 
> > > > It's happening significantly more lately.  Possibly because the
> > > > gmail
> > > > client helpfully hides quoted text.
> > > 
> > > I would also point to reviewers and maintainers who give a Rb/Ack
> > > tag:
> > > 1. somewhere at the top, without any footer like Best regards, and
> > > then
> > > quote entire patch, so I don't know shall I look for more comments
> > > after
> > > Rb/Ack?
> > > 
> > > 2. quote entire email and then add Rb/Ack, so I need to figure out
> > > whether there was something between the hundreds of lines of text
> > > or not.
> > 
> > Here we all are, brilliantly talented computer programmers who spend
> > our days making amazing fast digital devices do amazingly clever and
> > subtle things, inventing time-saving tools and processing vast
> > amounts of data without blinking, but for some reason we think the
> > task of skipping over a few thousand lines that all start with '> "
> > is too hard for us and that we should, in stead, complain to some
> > other human to convince them to make our life easier for us.
> > 
> > Does anyone else see the irony?
> 
> So if I'm a brilliantly talented driver, it's OK for other people to
> drive on the wrong side of the road because I should be able to avoid
> them?

No, we are programmers, not MUAs or drivers.  We program the car to
predict and avoid all other road users, no matter where they are.
That project might take a bit longer than fixing MUAs though.

> 
> The point being there are some situations where observing global
> etiquette is way more helpful than an individual solution, however
> talented the individual.

True, and we all use (some version of) English for exactly that reason.
But imagine how it would be if we could all have high quality
translation code built into our MUAs so that anyone in the world could
email us in their own language, and we could each read in our own
language.  That would be awesome.
It would be only slightly less awesome if we could all post with
whichever quoting style works for us, and we could all read emails
seeing whichever quoting style works for us.
Unfortunately we cannot yet translate top-posting to post-posting, but
we *can* translate verbose quoting to terse quoting.

NeilBrown

Re: KTODO automated TODO lists

[Randy Dunlap]

On 10/25/23 14:14, NeilBrown wrote:
> On Wed, 25 Oct 2023, Geert Uytterhoeven wrote:
>> Hi Neil,
>>
>> On Tue, Oct 24, 2023 at 11:29 PM NeilBrown <neilb@suse.de> wrote:
>>> On Wed, 25 Oct 2023, Krzysztof Kozlowski wrote:
>>>> On 23/10/2023 20:49, Andrew Morton wrote:
>>>>> On Thu, 19 Oct 2023 07:11:36 +0300 Dan Carpenter <dan.carpenter@linaro.org> wrote:
>>>>>> Yesterday someone on my lists just sent an email looking for kernel
>>>>>> tasks.
>>>>>
>>>>> Well here's a task: write a bot which follows the mailing lists and
>>>>> sends people nastygrams if one of their emails is more than 95%(?)
>>>>> quoted text.
>>>>>
>>>>> It's happening significantly more lately.  Possibly because the gmail
>>>>> client helpfully hides quoted text.
>>>>
>>>> I would also point to reviewers and maintainers who give a Rb/Ack tag:
>>>> 1. somewhere at the top, without any footer like Best regards, and then
>>>> quote entire patch, so I don't know shall I look for more comments after
>>>> Rb/Ack?
>>>>
>>>> 2. quote entire email and then add Rb/Ack, so I need to figure out
>>>> whether there was something between the hundreds of lines of text or not.
>>>
>>> Here we all are, brilliantly talented computer programmers who spend
>>> our days making amazing fast digital devices do amazingly clever and
>>> subtle things, inventing time-saving tools and processing vast amounts
>>> of data without blinking, but for some reason we think the task of
>>> skipping over a few thousand lines that all start with '> " is too hard
>>> for us and that we should, in stead, complain to some other human to
>>> convince them to make our life easier for us.
>>>
>>> Does anyone else see the irony?
>>
>> Please compare the numbers:
>>   1. 1 sender removes irrelevant parts,
>>   2. N receivers skip irrelevant parts.
> 
> That is one way to look at the numbers.
> Another is:
> 
>   12 - fix about a dozen MUAs to summaries quotes properly
>   12000 - fix an unknownable number of people to quote just exactly the
>           amount that their particular audience is going to want
> 
> and when it comes to fixing-code versus fixing-people, I know which this
> community is better at.
> 
> I guess there is also the option
> 
>    1 - fix vger.kernel.org to reject postings from people who don't
>         think and quote like "us", because we already have too many
>         contributor and want to block the heretics
> 
> This is really just a form of the "platform problem" which lwn.net has
> occasionally written about.  The "problem" is that we treat the platform
> (library code or other infrastructure) as fixed and develop ugly hacks
> in our own code to work around some shortcoming, rather the going into
> the platform and fixing it once for everyone there.

The problem AFAICT is that many (most?) of us expect a certain level of
etiquette but we are not seeing it in some posts.

-- 
~Randy

Re: KTODO automated TODO lists

[Steven Rostedt]

On Tue, 24 Oct 2023 20:47:28 -0700
"Paul E. McKenney" <paulmck@kernel.org> wrote:

> On Tue, Oct 24, 2023 at 06:05:17PM -0400, Steven Rostedt wrote:
> > On Wed, 25 Oct 2023 08:29:14 +1100
> > "NeilBrown" <neilb@suse.de> wrote:
> >   
> > > Here we all are, brilliantly talented computer programmers who spend
> > > our days making amazing fast digital devices do amazingly clever and
> > > subtle things, inventing time-saving tools and processing vast amounts
> > > of data without blinking, but for some reason we think the task of
> > > skipping over a few thousand lines that all start with '> " is too hard
> > > for us and that we should, in stead, complain to some other human to
> > > convince them to make our life easier for us.
> > > 
> > > Does anyone else see the irony?  
> > 
> > Did you also know that real-time developers are the most unpredictable?  
> 
> Are safety-critical programmers the most easy-going?
> 

No, they are the most accident prone ;-)

-- Steve

Re: KTODO automated TODO lists

[Dan Carpenter]

On Thu, Oct 26, 2023 at 08:14:25AM +1100, NeilBrown wrote:
> On Wed, 25 Oct 2023, Geert Uytterhoeven wrote:
> > 
> > Please compare the numbers:
> >   1. 1 sender removes irrelevant parts,
> >   2. N receivers skip irrelevant parts.
> 
> That is one way to look at the numbers.
> Another is:
> 
>   12 - fix about a dozen MUAs to summaries quotes properly
>   12000 - fix an unknownable number of people to quote just exactly the
>           amount that their particular audience is going to want
> 
> and when it comes to fixing-code versus fixing-people, I know which this
> community is better at.

We've historically been successful at enforcing LKML etiquette rules on
everyone.  This is just another rule to not quote the entire email when
you're replying to a patch.

If you're just adding a Reviewed-by tag then post some context but not
more than a page.

For a new driver, what I sometimes to is put a summary at the top.
"Thanks.  The only real bug is some missing error codes in probe.  I had
some other style nits as well.  See below for all the details."  I
normally write the email first and then chop out the "no comment"
functions at the end.  (Sometimes I chop out the no comment functions
at the begining and then I have to start over when I change my mind).

regards,
dan carpenter

Re: KTODO automated TODO lists

[Geert Uytterhoeven]

On Thu, Oct 26, 2023 at 6:29 AM Dan Carpenter <dan.carpenter@linaro.org> wrote:
> For a new driver, what I sometimes to is put a summary at the top.
> "Thanks.  The only real bug is some missing error codes in probe.  I had
> some other style nits as well.  See below for all the details."  I
> normally write the email first and then chop out the "no comment"
> functions at the end.  (Sometimes I chop out the no comment functions
> at the begining and then I have to start over when I change my mind).

Sounds very familiar (also the (...) part ;-)

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: KTODO automated TODO lists

[Miguel Ojeda]

On Tue, Oct 24, 2023 at 5:28 PM Konstantin Ryabitsev <mricon@kernel.org> wrote:
>
> FWIW, such page now exists:
> https://subspace.kernel.org/etiquette.html

Thanks Konstantin -- I have added a link to it in the Rust for Linux webpage.

I would suggest linking it at
https://docs.kernel.org/process/submitting-patches.html#use-trimmed-interleaved-replies-in-email-discussions
too.

Cheers,
Miguel

Re: KTODO automated TODO lists

[Steven Rostedt]

On Mon, 23 Oct 2023 17:38:15 -0600
"Gustavo A. R. Silva" <gustavo@embeddedor.com> wrote:

> On 10/18/23 22:11, Dan Carpenter wrote:
> > Yesterday someone on my lists just sent an email looking for kernel
> > tasks.  This was a university student in a kernel programming class.
> > We also have kernel-janitors and outreachy and those people are always
> > asking for small tasks.  
> 
> We have tons of issues waiting to be audited and fixed here:
> 
> https://scan.coverity.com/projects/linux-next-weekly-scan
> 
> You will never run out of fun. :) People just need to sign up.
> 
> That's really a great way to learn and gain experience across the whole
> kernel tree.
>

The difference between this and the KTODO is that the above is bugs that a
bot has discovered, right?

Although I agree that fixing bugs is a great way to learn the kernel, in
some cases people want to create a feature. At least that's a bit more
rewarding.

Currently, while working on adding a feature to the tracing ring buffer,
I've come across several bugs (that I fixed), but also a list of "nice to
haves".

That is, small feature enhancements that make the system better, that I
simply do not have the time to implement. This is where I think KTODO is
useful. If someone wants to add these enhancements, I'd be happy to help
them out (sparingly).

-- Steve