docutils-users

[Docutils-users] New version of rst.el

[Stefan M. <sm...@oe...>]

Hi Emacs users!

I committed a new version of `rst.el`_ to r6713. This version contains
major internal refactorings most of which are not visible for the
user. There is quite some test coverage meanwhile so refactoring
should introduce no new errors.

.. _rst.el: svn://svn.berlios.de/docutils/trunk/docutils/tools/editors/emacs/rst.el

These are the user visible changes hoping I didn't forget any:

Key bindings changed
====================

According to Emacs standards the old key bindings left no room for new
commands. I decided to rebind most keys to have two keys after the
leading C-c so there is room for new commands. Another point is to use
key bindings which are well known from other Emacs modes and therefore
cause less hassle to the experienced Emacser. When in reStructuredText
mode try C-c C-h for the new bindings.

Old bindings are deprecated. Most of them still work but give a
warning informing about the new key binding.

I know that this measure may cause pain. However, it was necessary to
make further development possible.

Font-locking
============

Section manipulation and font-locking share the same idea of the
section adornment hierarchy and recognize section adornment better.
Thus font-locking and section adornment is more consistent and
correct.

`jit-lock-mode` is now on by default. On slow machines this may cause
a slow down close to huge blocks of comments or literal text. I'm in
contact with the Emacs maintainer about this problem but it is a
general problem in `jit-lock-mode`. I'm interested in problem reports.
If it doesn't work properly for too many I'll change that again.

New list insertion command
==========================

Based on code from Wei-Wei Guo there is now `rst-insert-list` (C-c C-l
C-i) inserting a list item matching the current list.

Shift region command changed
============================

`rst-shift-region` now works like `indent-rigidly`. I.e. you can give
a positive or negative argument to control direction and amount of
reStructuredText indendation points you want to change indendation of
the region by. There is only one key binding (C-c C-r TAB).

`rst-shift-region` does not fill any more - this should be left to the
user. Variable `rst-shift-fill-region` is removed therefore.
`rst-shift-region` doesn't drop the region.

`rst-indent-width` has been renamed from `rst-shift-basic-offset` and
is customizable.

Ideas
=====

There is a `file for ideas`_. Suggestions for further improvement are
welcome.

.. _file for ideas: svn://svn.berlios.de/docutils/trunk/docutils/tools/editors/emacs/IDEAS.rst

Minor changes
=============

`rst-preferred-decorations` has been renamed to
`rst-preferred-adornments`. An `define-obsolete-variable-alias` is in
place so this should do no harm. `rst-preferred-decorations` has a
proper customization type.

Compilation commands use and remove temporary files fixing bug
#2912890.

Removed compatibility code for Emacs before V22.

`paragraph-separate` is set explicitly and `paragraph-start` is more
valid.

Bug #2972588 fixed as a side effect of refactorings.


						Grüße

						Stefan

[Docutils-users] New version of rst.el

[Stefan M. <st...@me...>]

Hi!

I just released a version V1.4.1 of rst.el:
http://sourceforge.net/p/docutils/code/HEAD/tree/trunk/docutils/tools/editors/emacs/rst.el

This is mainly a bug fix release with no notable feature changes. The
most important problem fixed came up with Emacs 24.4 which switched
`electric-indent-mode` globally on which broke `rst.el` awfully.
`rst.el` in the main Emacs development tree fixed the problem already.

I also improved the tutorial somewhat.


						Grüße

						Stefan

[Docutils-users] New version of rst.el

[Stefan M. <st...@me...>]

Hi!

I just released version V1.4.2 of rst.el:
http://sourceforge.net/p/docutils/code/HEAD/tree/trunk/docutils/tools/editors/emacs/rst.el

This is mainly a refactoring release with only few user visible
feature changes. I replaced the macros `rst-iterate-leftmost-...` by
the new function `rst-apply-indented-blocks`. I refactored
`rst-enumerate-region`, `rst-bullet-list-region` and
`rst-line-block-region` to use this new function.

During that I improved these commands a bit so they are more in line
with reStructuredText logic.

Thanks to Dominik who pointed out an endless loop in these functions
under certain circumstances. This was my starting point to refactor
all this code.


						Grüße

						Stefan

Re: [Docutils-users] New version of rst.el

[Dmitry S. <mi...@gm...>]

Hi Stefan,

Stefan Merten writes:
> Hi!
>
> I just released version V1.4.2 of rst.el:
>
http://sourceforge.net/p/docutils/code/HEAD/tree/trunk/docutils/tools/editors/emacs/rst.el

It would be very nice if you could look at
https://sourceforge.net/p/docutils/bugs/283/ (which I have forwarded from
Debian bug tracking system)?

--
Dmitry Shachnev

Re: [Docutils-users] New version of rst.el

[Stefan M. <st...@me...>]

Hi Dmitry!

3 days ago Dmitry Shachnev wrote:
> It would be very nice if you could look at
> https://sourceforge.net/p/docutils/bugs/283/ (which I have forwarded from
> Debian bug tracking system)?

Thanks for the reminder.

This is indeed not a bug but a feature request: Definition lists are
not supported yet. See the comment before `rst-line-tabs`::

  ;; FIXME: Must consider other tabs:
  ;;        * Line blocks
  ;;        * Definition lists
  ;;        * Option lists

Definition lists seem extra hard to me - but frankly I never gave them
a try.

What to do in such a case?


						Grüße

						Stefan

Re: [Docutils-users] New version of rst.el

[Dmitry S. <mi...@gm...>]

Hi Stefan!

Stefan Merten <stefan <at> merten-home.de> writes:
> Thanks for the reminder.
>
> This is indeed not a bug but a feature request: Definition lists are
> not supported yet. See the comment before `rst-line-tabs`::
>
>   ;; FIXME: Must consider other tabs:
>   ;;        * Line blocks
>   ;;        * Definition lists
>   ;;        * Option lists
>
> Definition lists seem extra hard to me - but frankly I never gave them
> a try.
>
> What to do in such a case?

[ Sorry for the late reply — actually I was not subscribed to this list so
missed your message. ]

If you think this is non-trivial to fix then you can probably ignore it (but
let's keep the bug open in case someone else decides to look at it).

I am not a user of Emacs or rst.el so this won't be me :)

--
Dmitry Shachnev

Re: [Docutils-users] New version of rst.el

[Ben F. <ben...@be...>]

Stefan Merten <sm...@oe...> writes:

> Key bindings changed
> ====================
>
> According to Emacs standards the old key bindings left no room for new
> commands. I decided to rebind most keys to have two keys after the
> leading C-c so there is room for new commands.

By my reading, that means most commands use three keys in total (leading
C-c plus two keys). Is that right?

If so, I don't see why that's been done. Isn't the point of a leading
C-c to make the key combinations not clash with Emacs built-ins? Surely
the C-c prefixed commands should be mostly two keys total, no?

-- 
 \     “I went to the cinema, it said ‘Adults: $5.00, Children $2.50’. |
  `\          So I said ‘Give me two boys and a girl.’” —Steven Wright |
_o__)                                                                  |
Ben Finney

Re: [Docutils-users] New version of rst.el

[Stefan M. <sm...@oe...>]

Hi Ben and all!

Yesterday Ben Finney wrote:
>> Key bindings changed
>> ====================
>> 
>> According to Emacs standards the old key bindings left no room for new
>> commands. I decided to rebind most keys to have two keys after the
>> leading C-c so there is room for new commands.
> 
> By my reading, that means most commands use three keys in total (leading
> C-c plus two keys). Is that right?

True.

> If so, I don't see why that's been done. Isn't the point of a leading
> C-c to make the key combinations not clash with Emacs built-ins? Surely
> the C-c prefixed commands should be mostly two keys total, no?

There is an Emacs internal standard that major modes should use only
certain key bindings:

  * Sequences consisting of `C-c' followed by a control character or a
  digit are reserved for major modes.

  -- (elisp)Key Binding Conventions

I'm trying to follow this convention. For major modes this means that
there are only ~36 possible key bindings with a single key. Since for
feature rich major modes like reST mode this is pretty few and leads
to meaningless and thus hard to remember key combinations there needs
to be another solution. Using two keys is the best I found so far.
Also this is not unusual - see for instance the history of key
bindings of MH-E.

I know rebinding keys has a high user impact. I hate it myself. But I
think there are enough reasons to do it and I do my best to make the
shift as smooth as possible. If someone comes up with a better
alternative which also complies with the Emacs conventions I'm happy
to implement this.

BTW: I'm also in the process of becoming a regular Emacs developer so
I can maintain rst.el in the main Emacs development as well. But so
far I tend to maintain the upstream version here since the feedback
here is much better.


						Grüße

						Stefan

Re: [Docutils-users] New version of rst.el

[Ben F. <ben...@be...>]

Stefan Merten <sm...@oe...> writes:

> I know rebinding keys has a high user impact. I hate it myself. But I
> think there are enough reasons to do it and I do my best to make the
> shift as smooth as possible. If someone comes up with a better
> alternative which also complies with the Emacs conventions I'm happy
> to implement this.

Well, as the person doing the work you have more of a say in how it gets
done than me. So, thank you for your effort and maintenance of the
useful rst-mode :-)

-- 
 \                “Stop — Drive sideways.” —detour sign, Kyushu, Japan |
  `\                                                                   |
_o__)                                                                  |
Ben Finney