Re: [PATCH] sequencer: support folding in rfc2822 footer

[Jeff King]

On Tue, Sep 06, 2016 at 04:30:24PM -0700, Jonathan Tan wrote:

> On 09/06/2016 03:08 PM, Jonathan Tan wrote:
> > On 09/02/2016 07:23 PM, Junio C Hamano wrote:
> > > A slightly related tangent.  An unconditionally good change you
> > > could make is to allow folding of in-body headers.  I.e. you can
> > > have e.g.
> > > 
> > > -- >8 --
> > > Subject: [PATCH] sequencer: support in-body headers that are
> > >  folded according to RFC2822 rules
> > > 
> > > The first paragraph after the above long title begins
> > > here...
> > > 
> > > in the body of the msssage, and I _think_ we do not fold it properly
> > > when applying such a patch.  We should, as that is something that
> > > appears in format-patch output (i.e. something Git itself produces,
> > > unlike the folded "footer").
> > 
> > OK, I'll take a look at this.
> 
> It turns out that Git seems to already do this, at least for Subject.

Right, because "Subject" is actually a real RFC 2822 header in the
generated email message. Not only do we expect things like mail readers
to handle this, but we _have_ to wrap at a certain point to meet the
standard[1].

I don't think any part of Git ever shunts "Subject" to an in-body
header, though I'd guess people do it manually all the time. 

> $ git format-patch HEAD^
> 0001-this-is-a-very-long-subject-to-test-line-wrapping-th.patch
> $ cat 0001-this-is-a-very-long-subject-to-test-line-wrapping-th.patch
> 
> Subject: [PATCH] this is a very long subject to test line wrapping this is a
>  very long subject to test line wrapping
> 

So the interesting bit is what happens with:

  git checkout master^
  git am 0001-*

and with:

  perl -lpe '
# Bump subject down to in-body header.
if (/^Subject:/) {
print "Subject: real subject";
print "";
}
  ' 0001-* >patch
  git checkout master^
  git am patch

It looks like we get the first one right, but not the second.

-Peff

[1] A careful reader may note that arbitrarily-long body lines,
including in-body headers and footers, may _also_ run afoul of
the body line-length limits. The "right" solution there is
probably quoted-printable, but it's ugly enough that I wouldn't do
so unless we see a real-world case where the line lengths are a
problem.

Re: [PATCH] sequencer: support folding in rfc2822 footer

[Jonathan Tan]

On 09/06/2016 03:08 PM, Jonathan Tan wrote:

On 09/02/2016 07:23 PM, Junio C Hamano wrote:

A slightly related tangent.  An unconditionally good change you
could make is to allow folding of in-body headers.  I.e. you can
have e.g.

-- >8 --
Subject: [PATCH] sequencer: support in-body headers that are
 folded according to RFC2822 rules

The first paragraph after the above long title begins
here...

in the body of the msssage, and I _think_ we do not fold it properly
when applying such a patch.  We should, as that is something that
appears in format-patch output (i.e. something Git itself produces,
unlike the folded "footer").


OK, I'll take a look at this.


It turns out that Git seems to already do this, at least for Subject. 
Transcript below:


$ echo one > file.txt
$ git add file.txt
$ git commit -m x
[master (root-commit) 2389483] x
 1 file changed, 1 insertion(+)
 create mode 100644 file.txt
$ echo two > file.txt
$ git commit -am 'this is a very long subject to test line wrapping this 
is a very long subject to test line wrapping'
[master ca86792] this is a very long subject to test line wrapping this 
is a very long subject to test line wrapping

 1 file changed, 1 insertion(+), 1 deletion(-)
$ git format-patch HEAD^
0001-this-is-a-very-long-subject-to-test-line-wrapping-th.patch
$ cat 0001-this-is-a-very-long-subject-to-test-line-wrapping-th.patch

Subject: [PATCH] this is a very long subject to test line wrapping this is a
 very long subject to test line wrapping

Re: [PATCH] sequencer: support folding in rfc2822 footer

[Jonathan Tan]

On 09/02/2016 07:23 PM, Junio C Hamano wrote:

Jonathan Tan  writes:


Sample-field: multiple-line field body
 that causes a blank line below


I am not sure this is unconditionally good, or may cause problems to
those with workflows you did not consider when you wrote this patch.

Not being too lenient here historically has been a deliberate
decision to avoid misidentification of non "footers".  Does Git
itself produce some folded footer line?  If Git itself produced such
folded lines, I'd be a lot more receptive to this change, but I do
not think that is the case here.


I don't think Git produces any folded lines, but folded lines do appear 
in footers in projects that use Git. For example, some Android commits 
have multi-line "Test:" fields (example, [1]) and some Linux commits 
have multi-line "Tested-by:" fields (example, [2]).


Taking the Android commit as an example, this would mean that 
cherrypicking that commit would create a whole new footer, and tripping 
up tools (for example, Gerrit, which looks for "Change-Id:" in the last 
paragraph). But this would not happen if "Test:" was single-line instead 
of multi-line - which seems inconsistent.


[1] 
https://android.googlesource.com/platform/frameworks/base/+/4c5281862f750cbc9d7355a07ef1a5545b9b3523
[2] 
https://kernel.googlesource.com/pub/scm/linux/kernel/git/stable/linux-stable/+/69f92f67b68ab7028ffe15f0eea76b59f8859383



A slightly related tangent.  An unconditionally good change you
could make is to allow folding of in-body headers.  I.e. you can
have e.g.

-- >8 --
Subject: [PATCH] sequencer: support in-body headers that are
 folded according to RFC2822 rules

The first paragraph after the above long title begins
here...

in the body of the msssage, and I _think_ we do not fold it properly
when applying such a patch.  We should, as that is something that
appears in format-patch output (i.e. something Git itself produces,
unlike the folded "footer").


OK, I'll take a look at this.

Re: [PATCH] sequencer: support folding in rfc2822 footer

[Junio C Hamano]

Jonathan Tan  writes:

> Sample-field: multiple-line field body
>  that causes a blank line below

I am not sure this is unconditionally good, or may cause problems to
those with workflows you did not consider when you wrote this patch.

Not being too lenient here historically has been a deliberate
decision to avoid misidentification of non "footers".  Does Git
itself produce some folded footer line?  If Git itself produced such
folded lines, I'd be a lot more receptive to this change, but I do
not think that is the case here.

A slightly related tangent.  An unconditionally good change you
could make is to allow folding of in-body headers.  I.e. you can
have e.g.

-- >8 --
Subject: [PATCH] sequencer: support in-body headers that are
 folded according to RFC2822 rules

The first paragraph after the above long title begins
here...

in the body of the msssage, and I _think_ we do not fold it properly
when applying such a patch.  We should, as that is something that
appears in format-patch output (i.e. something Git itself produces,
unlike the folded "footer").