b4 ty : wrong patch number when cherrypicking

b4 ty : wrong patch number when cherrypicking

[Alexandre Belloni]

Hello,

I cherrypicked patch 4 and 5 of
https://lore.kernel.org/linux-amlogic/20201005203451.9985-1-post@lespocky.de/
using:

b4 am -p linux-arm-kernel -c -t -s -l -P 4-5 -o - 20201005203451.9985-5-post@lespocky.de

After applying to my tree, b4 ty generated the message I sent (without
checking, my bad):

https://lore.kernel.org/linux-amlogic/160391135997.385141.5554228268638639718.b4-ty@bootlin.com/

Where it had 1/2 and 2/2 instead of 04/12 and 05/12.

Is this too awful to fix? :)

Regards,

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

Re: b4 ty : wrong patch number when cherrypicking

[Konstantin Ryabitsev]

On Fri, Oct 30, 2020 at 06:37:44PM +0100, Alexandre Belloni wrote:
> I cherrypicked patch 4 and 5 of
> https://lore.kernel.org/linux-amlogic/20201005203451.9985-1-post@lespocky.de/
> using:
> 
> b4 am -p linux-arm-kernel -c -t -s -l -P 4-5 -o - 20201005203451.9985-5-post@lespocky.de
> 
> After applying to my tree, b4 ty generated the message I sent (without
> checking, my bad):
> 
> https://lore.kernel.org/linux-amlogic/160391135997.385141.5554228268638639718.b4-ty@bootlin.com/
> 
> Where it had 1/2 and 2/2 instead of 04/12 and 05/12.
> 
> Is this too awful to fix? :)

It shouldn't be -- I have a similar request in the following thread:
https://lore.kernel.org/r/20201003193137.z2bpwzlz5a66kkex@chatter.i7.local

-K

Re: [tools] b4 ty : wrong patch number when cherrypicking

[Konstantin Ryabitsev]

On Fri, Oct 30, 2020 at 02:05:30PM -0400, Konstantin Ryabitsev via linux.kernel.org wrote:
> On Fri, Oct 30, 2020 at 06:37:44PM +0100, Alexandre Belloni wrote:
> > I cherrypicked patch 4 and 5 of
> > https://lore.kernel.org/linux-amlogic/20201005203451.9985-1-post@lespocky.de/
> > using:
> > 
> > b4 am -p linux-arm-kernel -c -t -s -l -P 4-5 -o - 20201005203451.9985-5-post@lespocky.de
> > 
> > After applying to my tree, b4 ty generated the message I sent (without
> > checking, my bad):
> > 
> > https://lore.kernel.org/linux-amlogic/160391135997.385141.5554228268638639718.b4-ty@bootlin.com/
> > 
> > Where it had 1/2 and 2/2 instead of 04/12 and 05/12.
> > 
> > Is this too awful to fix? :)
> 
> It shouldn't be -- I have a similar request in the following thread:
> https://lore.kernel.org/r/20201003193137.z2bpwzlz5a66kkex@chatter.i7.local

There is a fix in the latest master that more properly handles these 
situations:

- we record actual patch prefixes when we start tracking a series, so 
  when we send a thank you report, the actual counters will be shown
- when we send a reply for a subset of the series, we will now insert a 
  "(subset)" into the subject to more clearly indicate to the submitter 
  that the series only got applied partially

Note, that this will only start working for *new* series you grab with 
"b4 am" -- anything you've grab previously will exhibit previous 
behaviour.

-K

Re: [kernel.org users] [tools] b4 ty : wrong patch number when cherrypicking

[Rob Herring]

On Tue, Nov 24, 2020 at 11:23 AM Konstantin Ryabitsev
<konstantin@linuxfoundation.org> wrote:
>
> On Fri, Oct 30, 2020 at 02:05:30PM -0400, Konstantin Ryabitsev via linux.kernel.org wrote:
> > On Fri, Oct 30, 2020 at 06:37:44PM +0100, Alexandre Belloni wrote:
> > > I cherrypicked patch 4 and 5 of
> > > https://lore.kernel.org/linux-amlogic/20201005203451.9985-1-post@lespocky.de/
> > > using:
> > >
> > > b4 am -p linux-arm-kernel -c -t -s -l -P 4-5 -o - 20201005203451.9985-5-post@lespocky.de
> > >
> > > After applying to my tree, b4 ty generated the message I sent (without
> > > checking, my bad):
> > >
> > > https://lore.kernel.org/linux-amlogic/160391135997.385141.5554228268638639718.b4-ty@bootlin.com/
> > >
> > > Where it had 1/2 and 2/2 instead of 04/12 and 05/12.
> > >
> > > Is this too awful to fix? :)
> >
> > It shouldn't be -- I have a similar request in the following thread:
> > https://lore.kernel.org/r/20201003193137.z2bpwzlz5a66kkex@chatter.i7.local
>
> There is a fix in the latest master that more properly handles these
> situations:
>
> - we record actual patch prefixes when we start tracking a series, so
>   when we send a thank you report, the actual counters will be shown
> - when we send a reply for a subset of the series, we will now insert a
>   "(subset)" into the subject to more clearly indicate to the submitter
>   that the series only got applied partially

Could we not modify subjects as that breaks gmail 'threading'. Yes,
gmail is broken, but there's more than a few users...

Rob

Re: [kernel.org users] [tools] b4 ty : wrong patch number when cherrypicking

[Konstantin Ryabitsev]

On Thu, Dec 03, 2020 at 08:03:33AM -0700, Rob Herring wrote:
> > > It shouldn't be -- I have a similar request in the following 
> > > thread:
> > > https://lore.kernel.org/r/20201003193137.z2bpwzlz5a66kkex@chatter.i7.local
> >
> > There is a fix in the latest master that more properly handles these
> > situations:
> >
> > - we record actual patch prefixes when we start tracking a series, so
> >   when we send a thank you report, the actual counters will be shown
> > - when we send a reply for a subset of the series, we will now insert a
> >   "(subset)" into the subject to more clearly indicate to the submitter
> >   that the series only got applied partially
> 
> Could we not modify subjects as that breaks gmail 'threading'. Yes,
> gmail is broken, but there's more than a few users...

Well, the goal is to make it extremely obvious for the submitter that 
only a subset of their patch series got applied -- so I would argue that
Gmail thread-breaking only helps here.

Overall, this would be an very rare situation, so I doubt anyone would 
care enough to get annoyed.

-K

Re: [kernel.org users] [tools] b4 ty : wrong patch number when cherrypicking

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 1154 bytes --]

On Tue, Dec 08, 2020 at 08:59:26AM -0500, Konstantin Ryabitsev wrote:
> On Thu, Dec 03, 2020 at 08:03:33AM -0700, Rob Herring wrote:

> > > - when we send a reply for a subset of the series, we will now insert a
> > >   "(subset)" into the subject to more clearly indicate to the submitter
> > >   that the series only got applied partially

> > Could we not modify subjects as that breaks gmail 'threading'. Yes,
> > gmail is broken, but there's more than a few users...

> Well, the goal is to make it extremely obvious for the submitter that 
> only a subset of their patch series got applied -- so I would argue that
> Gmail thread-breaking only helps here.

> Overall, this would be an very rare situation, so I doubt anyone would 
> care enough to get annoyed.

This is actually a pretty common situation - the two most frequent
situations are people doing cleanup work who decide to thread a bunch of
patches to different subsystems with no interdependencies together and
people adding new drivers that cross multiple subsystems (eg, a driver
plus DTS updates adding it to some platform or a MFD with several
subfunctions).

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 499 bytes --]

Re: [kernel.org users] [tools] b4 ty : wrong patch number when cherrypicking

[Konstantin Ryabitsev]

On Tue, Dec 08, 2020 at 02:03:48PM +0000, Mark Brown wrote:
> > Well, the goal is to make it extremely obvious for the submitter 
> > that only a subset of their patch series got applied -- so I would 
> > argue that
> > Gmail thread-breaking only helps here.
> 
> > Overall, this would be an very rare situation, so I doubt anyone would 
> > care enough to get annoyed.
> 
> This is actually a pretty common situation - the two most frequent
> situations are people doing cleanup work who decide to thread a bunch of
> patches to different subsystems with no interdependencies together and
> people adding new drivers that cross multiple subsystems (eg, a driver
> plus DTS updates adding it to some platform or a MFD with several
> subfunctions).

Okay, good to know. I'm still sticking with "(subset)" in the subject, 
as this hopefully avoids more problems than it causes.

-K

Re: [kernel.org users] [tools] b4 ty : wrong patch number when cherrypicking

[Rob Herring]

On Tue, Dec 8, 2020 at 7:59 AM Konstantin Ryabitsev
<konstantin@linuxfoundation.org> wrote:
>
> On Thu, Dec 03, 2020 at 08:03:33AM -0700, Rob Herring wrote:
> > > > It shouldn't be -- I have a similar request in the following
> > > > thread:
> > > > https://lore.kernel.org/r/20201003193137.z2bpwzlz5a66kkex@chatter.i7.local
> > >
> > > There is a fix in the latest master that more properly handles these
> > > situations:
> > >
> > > - we record actual patch prefixes when we start tracking a series, so
> > >   when we send a thank you report, the actual counters will be shown
> > > - when we send a reply for a subset of the series, we will now insert a
> > >   "(subset)" into the subject to more clearly indicate to the submitter
> > >   that the series only got applied partially
> >
> > Could we not modify subjects as that breaks gmail 'threading'. Yes,
> > gmail is broken, but there's more than a few users...
>
> Well, the goal is to make it extremely obvious for the submitter that
> only a subset of their patch series got applied -- so I would argue that
> Gmail thread-breaking only helps here.

If you manually label things in gmail (or just look at 'Sent'),
replies don't get the label (and in turn aren't in the corresponding
IMAP folder). But they will still show up the gmail thread unless the
subject is changed. So the reply could just be missed completely
rather than be unclear to people that don't read the body.

> Overall, this would be an very rare situation, so I doubt anyone would
> care enough to get annoyed.

I'm annoyed, hence why I raised it. But if no one else cares, then
fine. I'm using mutt now in most situations I care about. I was more
concerned with less active submitters only using gmail.

Rob