Backporting changes into emacs-23 prior to pretest

Backporting changes into emacs-23 prior to pretest

[Chong Yidong]

In preparation for the Emacs 23.3 pretest, I'd like to ask everyone if
there are any specific parts of the trunk you'd like to see backported.
The 23.3 release will be mostly bugfixes, so significant new features
will not be considered.  However, updated "stable" versions of exsiting
modes and packages should be acceptable.


(One set of changes that I think would be good to backport are the
improvements to VC since emacs-23 was branched.  However, this is
complicated by the fact that the Lisp files have been moved into a vc/
directory in the trunk, so I'm not sure how to go about merging them.
Does anyone have a suggestion?)

Re: Backporting changes into emacs-23 prior to pretest

[Dan Nicolaescu]

Chong Yidong <cyd@stupidchicken.com> writes:

> In preparation for the Emacs 23.3 pretest, I'd like to ask everyone if
> there are any specific parts of the trunk you'd like to see backported.
> The 23.3 release will be mostly bugfixes, so significant new features
> will not be considered.  However, updated "stable" versions of exsiting
> modes and packages should be acceptable.

How about this one:
   ** Directory local variables can apply to file-less buffers.
   For example, adding "(diff-mode . ((mode . whitespace)))" to your
   .dir-locals.el file, will turn on `whitespace-mode' for *vc-diff* buffers.
The code change itself is small, and only touches one or two
functions.

How about the changes to make major modes derive from `prog-mode'?
[This is not quite finished, but it won't be too hard to finish it...]

> (One set of changes that I think would be good to backport are the
> improvements to VC since emacs-23 was branched.  However, this is
> complicated by the fact that the Lisp files have been moved into a vc/
> directory in the trunk, so I'm not sure how to go about merging them.
> Does anyone have a suggestion?)

Copy the files from trunk over to the branch.  AFAIR there's nothing
that depends on changes to the core, so just copying should be fine.

Re: Backporting changes into emacs-23 prior to pretest

[Ulrich Mueller]

>>>>> On Sun, 24 Oct 2010, Chong Yidong wrote:

> In preparation for the Emacs 23.3 pretest, I'd like to ask everyone
> if there are any specific parts of the trunk you'd like to see
> backported. The 23.3 release will be mostly bugfixes, so significant
> new features will not be considered. However, updated "stable"
> versions of exsiting modes and packages should be acceptable.

Hi,
I'm not sure if this qualifies, but I'd like to suggest that the
--with-crt-dir configure option is backported. It fixes Bug 5655 [1],
and in Gentoo we apply it to emacs-23.2 (in our stable tree) since
May, without any problems.

Corresponding commits are 100020 to 100022 [2]. (I don't remember any
more if it applied to 23.2 without conflicts. So just in case, a
backported patch for 23.2 is here: [3])

Ulrich

[1] <http://debbugs.gnu.org/5655>
[2] <http://bazaar.launchpad.net/~vcs-imports/emacs/trunk/revision/100020>
    <http://bazaar.launchpad.net/~vcs-imports/emacs/trunk/revision/100021>
    <http://bazaar.launchpad.net/~vcs-imports/emacs/trunk/revision/100022>
[3] <http://sources.gentoo.org/cgi-bin/viewvc.cgi/gentoo/src/patchsets/emacs/23.2/01_all_crt-dir.patch>

Re: Backporting changes into emacs-23 prior to pretest

[Chong Yidong]

Dan Nicolaescu <dann@gnu.org> writes:

> How about this one:
>    ** Directory local variables can apply to file-less buffers.
>    For example, adding "(diff-mode . ((mode . whitespace)))" to your
>    .dir-locals.el file, will turn on `whitespace-mode' for *vc-diff* buffers.
> The code change itself is small, and only touches one or two
> functions.

I don't think this is a good candidate for backporting, but if you feel
strongly about it, we can discuss more.

> How about the changes to make major modes derive from `prog-mode'?
> [This is not quite finished, but it won't be too hard to finish it...]

These is mostly a cleanup, right?  I don't think it's worth backporting
either.

>> (One set of changes that I think would be good to backport are the
>> improvements to VC since emacs-23 was branched.  However, this is
>> complicated by the fact that the Lisp files have been moved into a vc/
>> directory in the trunk, so I'm not sure how to go about merging them.
>> Does anyone have a suggestion?)
>
> Copy the files from trunk over to the branch.  AFAIR there's nothing
> that depends on changes to the core, so just copying should be fine.

OK, I've done this.  Thanks.

Re: Backporting changes into emacs-23 prior to pretest

[Chong Yidong]

Ulrich Mueller <ulm@gentoo.org> writes:

> I'm not sure if this qualifies, but I'd like to suggest that the
> --with-crt-dir configure option is backported. It fixes Bug 5655 [1],
> and in Gentoo we apply it to emacs-23.2 (in our stable tree) since
> May, without any problems.

This seems reasonable.  I've backported this to the branch.  There were
later developments in Bug#5655, but my understanding is that they are
not strictly necessary to fix the bug; is this correct?

Re: Backporting changes into emacs-23 prior to pretest

[Ulrich Mueller]

>>>>> On Wed, 03 Nov 2010, Chong Yidong wrote:

> Ulrich Mueller <ulm@gentoo.org> writes:
>> I'm not sure if this qualifies, but I'd like to suggest that the
>> --with-crt-dir configure option is backported. It fixes Bug 5655 [1],
>> and in Gentoo we apply it to emacs-23.2 (in our stable tree) since
>> May, without any problems.

> This seems reasonable.  I've backported this to the branch.  There were
> later developments in Bug#5655, but my understanding is that they are
> not strictly necessary to fix the bug; is this correct?

Yes, the commits mentioned in my previous message should be enough.
AFAICS, that's up to (and including) message #13 in the bug.

Ulrich

Re: Backporting changes into emacs-23 prior to pretest

[Dan Nicolaescu]

Chong Yidong <cyd@stupidchicken.com> writes:

> Dan Nicolaescu <dann@gnu.org> writes:
>
>> How about this one:
>>    ** Directory local variables can apply to file-less buffers.
>>    For example, adding "(diff-mode . ((mode . whitespace)))" to your
>>    .dir-locals.el file, will turn on `whitespace-mode' for *vc-diff* buffers.
>> The code change itself is small, and only touches one or two
>> functions.
>
> I don't think this is a good candidate for backporting, but if you feel
> strongly about it, we can discuss more.

I think this is important.  It completes what .dir-locals.el can do.
Without it, you cannot use .dir-locals.el for any of the *vc* buffers.
The line I cite above is used to turn on whitespace-mode in *vc diff*.

>
>> How about the changes to make major modes derive from `prog-mode'?
>> [This is not quite finished, but it won't be too hard to finish it...]
>
> These is mostly a cleanup, right?  I don't think it's worth backporting
> either.

It's a new feature.  If all programming modes are derived from
`prog-mode', then one can do:
(add-hook 'prog-mode 'flyspell-prog-mode)
to turn on the fly spell checking for comments and strings in all programming modes.
This is not complete on the trunk either.  So if you have an
aggressive schedule to get the release out, we can let this one go.

Re: Backporting changes into emacs-23 prior to pretest

[Stefan Monnier]

>>> How about the changes to make major modes derive from `prog-mode'?
>>> [This is not quite finished, but it won't be too hard to finish it...]
>> These is mostly a cleanup, right?  I don't think it's worth backporting
>> either.
> It's a new feature.  If all programming modes are derived from
> `prog-mode', then one can do:
> (add-hook 'prog-mode 'flyspell-prog-mode)
> to turn on the fly spell checking for comments and strings in all
> programming modes.  This is not complete on the trunk either.  So if
> you have an aggressive schedule to get the release out, we can let
> this one go.

It's just a new feature, which has been missing for many years and yet
hasn't generated too much noise over those years, so it's not
particularly pressing.


        Stefan