Re: master 2d7f4d9 12/15: ; Merge from origin/emacs-25

Re: master 2d7f4d9 12/15: ; Merge from origin/emacs-25

[Michael Albinus]

John Wiegley <johnw@gnu.org> writes:

>     The following commits were skipped:
>     
>     4e62197 Introduce check-expensive tests.

I've merged this to the master branch manually.

Best regards, Michael.

Re: master 2d7f4d9 12/15: ; Merge from origin/emacs-25

[Eli Zaretskii]

> From: Michael Albinus <michael.albinus@gmx.de>
> Date: Tue, 12 Jan 2016 10:48:07 +0100
> 
> John Wiegley <johnw@gnu.org> writes:
> 
> >     The following commits were skipped:
> >     
> >     4e62197 Introduce check-expensive tests.
> 
> I've merged this to the master branch manually.

John, why were those skipped?  Did you bump into prblems trying to
merge them?  Does gitmerge.el need some work?

It's not a catastrophe to have to make several merges by hand, but I
think we should in general avoid that, as doing too many of those is
error-prone.

Re: master 2d7f4d9 12/15: ; Merge from origin/emacs-25

[John Wiegley]

>>>>> Eli Zaretskii <eliz@gnu.org> writes:

> John, why were those skipped? Did you bump into prblems trying to merge
> them? Does gitmerge.el need some work?

I did run into a conflict, where master had deleted a Makefile target, and
emacs-25 had added a target. I had thought incorrectly that master was right.

This will continue to be a danger while I'm still not as familiar with the
code as I need to be to know the right resolutions. What I can do is consult
first with the commit authors when I run into conflicts that are not obvious.

> It's not a catastrophe to have to make several merges by hand, but I think
> we should in general avoid that, as doing too many of those is error-prone.

That should be the only misfire on this merge. The other changes were simpler,
and there were few conflicts.

-- 
John Wiegley                  GPG fingerprint = 4710 CF98 AF9B 327B B80F
http://newartisans.com                          60E1 46C4 BD1A 7AC1 4BA2

Re: master 70f3e4e 08/15: ; Merge from origin/emacs-25

[Glenn Morris]

John Wiegley wrote:

> branch: master
> commit 70f3e4e61fc4433ca8ac4517d247df7e7af7b248
> Merge: 4b739f7 5213ded
> Author: John Wiegley <johnw@newartisans.com>
> Commit: John Wiegley <johnw@newartisans.com>
>
>     ; Merge from origin/emacs-25
>     
>     The following commit was skipped:
>     
>     5213ded Refactor mml-smime.el, mml1991.el, mml2015.el

Why did you skip this?

In general, almost every non-trivial "skip" in this merge looks wrong to
me.

Re: master 70f3e4e 08/15: ; Merge from origin/emacs-25

[John Wiegley]

>>>>> Glenn Morris <rgm@gnu.org> writes:

> In general, almost every non-trivial "skip" in this merge looks wrong to me.

I accepted the skips from gitmerge. Looks like that is not the right thing to
do after all.

What is the right way to cherry-pick over the skipped commits that shouldn't
have been?

-- 
John Wiegley                  GPG fingerprint = 4710 CF98 AF9B 327B B80F
http://newartisans.com                          60E1 46C4 BD1A 7AC1 4BA2

Re: master 70f3e4e 08/15: ; Merge from origin/emacs-25

[Glenn Morris]

Oh, I guess it is because the original commit message happened to
contain the word "merge", and thus match gitmerge-skip-regexp. IIRC,
gitmerge should ask you whether to skip a given commit, not just
unconditionally skip it. Some judgement is required when deciding to
answer yes or no to that prompt. The gitmerge regexp is biased to match
more rather than less.

Re: master 70f3e4e 08/15: ; Merge from origin/emacs-25

[John Wiegley]

[-- Attachment #1: Type: text/plain, Size: 901 bytes --]

>>>>> Glenn Morris <rgm@gnu.org> writes:

> In general, almost every non-trivial "skip" in this merge looks wrong to me.

Here is what was skipped:

1  31a97ac  Michael Albinus     :  Merge branch 'emacs-25' of git.sv.gnu.org:...
2  e79b06e  Paul Eggert         :  Avoid stdio in SIGINT handler
3  5213ded  Jens Lechtenboerger :  Refactor mml-smime.el, mml1991.el, mml2015.el
4  372d00a  Glenn Morris        :  ; Auto-commit of loaddefs files.
5  9825305  Paul Eggert         :  Merge from gnulib
6  30f4a89  YAMAMOTO Mitsuharu  :  Move variables to inner loop, preparing for...

I take it #1 and #4 are valid skips, but the other four should be
cherry-picked into master? Do we have a special command or annotation for
doing so?

-- 
John Wiegley                  GPG fingerprint = 4710 CF98 AF9B 327B B80F
http://newartisans.com                          60E1 46C4 BD1A 7AC1 4BA2

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 629 bytes --]

Re: master 70f3e4e 08/15: ; Merge from origin/emacs-25

[John Wiegley]

>>>>> Glenn Morris <rgm@gnu.org> writes:

> Oh, I guess it is because the original commit message happened to contain
> the word "merge", and thus match gitmerge-skip-regexp. IIRC, gitmerge should
> ask you whether to skip a given commit, not just unconditionally skip it.
> Some judgement is required when deciding to answer yes or no to that prompt.
> The gitmerge regexp is biased to match more rather than less.

This was my mistake, in thinking gitmerge was DTRT. I'll review each skip
manually from now on.

-- 
John Wiegley                  GPG fingerprint = 4710 CF98 AF9B 327B B80F
http://newartisans.com                          60E1 46C4 BD1A 7AC1 4BA2

Re: master 70f3e4e 08/15: ; Merge from origin/emacs-25

[Paul Eggert]

On 01/12/2016 09:11 AM, John Wiegley wrote:
> 1  31a97ac  Michael Albinus     :  Merge branch 'emacs-25' of git.sv.gnu.org:...
> 2  e79b06e  Paul Eggert         :  Avoid stdio in SIGINT handler
> 3  5213ded  Jens Lechtenboerger :  Refactor mml-smime.el, mml1991.el, mml2015.el
> 4  372d00a  Glenn Morris        :  ; Auto-commit of loaddefs files.
> 5  9825305  Paul Eggert         :  Merge from gnulib
> 6  30f4a89  YAMAMOTO Mitsuharu  :  Move variables to inner loop, preparing for...
>
> I take it #1 and #4 are valid skips, but the other four should be
> cherry-picked into master? Do we have a special command or annotation for
> doing so?

I don't think so, no. I would just install the patches into master. 
(It's not likely that anyone will want to backport them into emacs-25. :-)

Re: master 70f3e4e 08/15: ; Merge from origin/emacs-25

[Phillip Lord]

Paul Eggert <eggert@cs.ucla.edu> writes:

> On 01/12/2016 09:11 AM, John Wiegley wrote:
>> 1  31a97ac  Michael Albinus     :  Merge branch 'emacs-25' of git.sv.gnu.org:...
>> 2  e79b06e  Paul Eggert         :  Avoid stdio in SIGINT handler
>> 3  5213ded  Jens Lechtenboerger :  Refactor mml-smime.el, mml1991.el, mml2015.el
>> 4  372d00a  Glenn Morris        :  ; Auto-commit of loaddefs files.
>> 5  9825305  Paul Eggert         :  Merge from gnulib
>> 6  30f4a89  YAMAMOTO Mitsuharu  :  Move variables to inner loop, preparing for...
>>
>> I take it #1 and #4 are valid skips, but the other four should be
>> cherry-picked into master? Do we have a special command or annotation for
>> doing so?
>
> I don't think so, no. I would just install the patches into master. (It's not
> likely that anyone will want to backport them into emacs-25. :-)


I think 372d00a should validly be skipped. The loaddefs files shouldn't
need auto-committing on master.

Phil

Re: master 70f3e4e 08/15: ; Merge from origin/emacs-25

[Glenn Morris]

Is anyone going to fix this merge?

Re: master 70f3e4e 08/15: ; Merge from origin/emacs-25

[Paul Eggert]

Glenn Morris wrote:
> Is anyone going to fix this merge?

What are the problems with it? (I have not been following the 'master' branch 
lately.)

Re: master 70f3e4e 08/15: ; Merge from origin/emacs-25

[John Wiegley]

>>>>> Glenn Morris <rgm@gnu.org> writes:

> Is anyone going to fix this merge?

Yes, it was on my list for the weekend, but will happen today.

-- 
John Wiegley                  GPG fingerprint = 4710 CF98 AF9B 327B B80F
http://newartisans.com                          60E1 46C4 BD1A 7AC1 4BA2

Re: master 70f3e4e 08/15: ; Merge from origin/emacs-25

[John Wiegley]

>>>>> Paul Eggert <eggert@cs.ucla.edu> writes:

> On 01/12/2016 09:11 AM, John Wiegley wrote:
> > 1  31a97ac  Michael Albinus     :  Merge branch 'emacs-25' of git.sv.gnu.org:...
> > 2  e79b06e  Paul Eggert         :  Avoid stdio in SIGINT handler
> > 3  5213ded  Jens Lechtenboerger :  Refactor mml-smime.el, mml1991.el, mml2015.el
> > 4  372d00a  Glenn Morris        :  ; Auto-commit of loaddefs files.
> > 5  9825305  Paul Eggert         :  Merge from gnulib
> > 6  30f4a89  YAMAMOTO Mitsuharu  :  Move variables to inner loop, preparing for...
> >
> > I take it #1 and #4 are valid skips, but the other four should be
> > cherry-picked into master? Do we have a special command or annotation for
> > doing so?
> 
> I don't think so, no. I would just install the patches into master. (It's not
> likely that anyone will want to backport them into emacs-25. :-)

Just a note that the four commits above have been cherry picked to master, and
another merge from emacs-25 has been performed.

-- 
John Wiegley                  GPG fingerprint = 4710 CF98 AF9B 327B B80F
http://newartisans.com                          60E1 46C4 BD1A 7AC1 4BA2