Using punctuation in abbrev

Using punctuation in abbrev

[Aurélien Aptel]

Hi,

I want to substitute $-> with a unicode rightwards arrow (U+2192  →).
Since abbrev defaults to backward-word to extract the word before
point, I've used the :regexp property of an abbrev-table to define
what to extract.

From define-abbrev-table documentation:

  - `:regexp' is a regular expression that specifies how to extract the
    name of the abbrev before point.  The submatch 1 is treated
    as the potential name of an abbrev.  If :regexp is nil, the default
    behavior uses `backward-word' and `forward-word' to extract the name
    of the abbrev, which can therefore only be a single word.

Here's what I put in my init file:

  (define-abbrev-table 'global-abbrev-table '(
                                              ("$->" "→")
                                              ("$=>" "⇒")
                                              ("$foo" "X")
                                              )
    "custom abbrev table"
    :regexp (rx (or bos bol space) (group (+ (not space)))))
  (setq save-abbrevs nil)
  (setq-default abbrev-mode t)

But only the $foo substitution works and I can't figure out why.
The regex seems to be correct:

(let ((rx (rx (or bos bol space) (group (+ (not space)))))
      (slist '(" $->" "\n$=>" " $foo")))
  (mapcar (lambda (s) (string-match-p rx s)) slist))

=> (0 0 0)

RE: Using punctuation in abbrev

[Drew Adams]

> I want to substitute $-> with a unicode rightwards arrow (U+2192  →).

1. Does that mean that you want to type `$->' and use `expand-abbrev'
each time to insert a Unicode right arrow?  If so, and if you don't
really care whether you use abbrev, consider binding insertion of a
right arrow char to a key sequence instead.

2. On the other hand, if you already have lots of `$->' occurrences in
your text and you want to change them to right arrows, consider using
`query-replace' or `replace-string' instead.

For #1, library ucs-cmds.el can help.
http://www.emacswiki.org/emacs-en/download/ucs-cmds.el
http://www.emacswiki.org/emacs/UnicodeEncoding

Command `ucsc-insert' is a replacement for `insert-char'.

You can remap the keys bound to `insert-char' to `uscs-insert', if
you like: (define-key global-map [remap insert-char] 'ucsc-insert)

Unless you give `ucsc-insert' a negative prefix arg it does the same
thing as `insert-char'.  With a negative prefix arg it creates a
command that inserts the Unicode character you choose.

You can provide the character by hex or by name (with completion).
In this case, that means 2192 or "RIGHT ARROW".

Create command `right-arrow', which inserts a right arrow char (U+2192):

M-- M-x ucsc-insert 2192 RET

Bind the command to a key (e.g., <f9>):

(global-set-key [f9] 'right-arrow)

You can use a numeric prefix arg to insert multiple right arrows:
`M-8 <f9>' inserts eight right arrows.

Re: Using punctuation in abbrev

[Aurélien Aptel]

On Sat, Jun 1, 2013 at 8:42 PM, Drew Adams <drew.adams@oracle.com> wrote:
>> I want to substitute $-> with a unicode rightwards arrow (U+2192  →).
>
> 1. Does that mean that you want to type `$->' and use `expand-abbrev'
> each time to insert a Unicode right arrow?  If so, and if you don't
> really care whether you use abbrev, consider binding insertion of a
> right arrow char to a key sequence instead.

The abbrev minor mode is active in all my buffers and it doesn't have
to be triggered by a specific key, it's automatically substituted as I
type. I plan on having a lot of easy-to-remember abbreviations for
unicode chars I regularly use, all prefixed by $ or some other less
used character so that it only triggers when I want. Calling
insert-char is too cumbersome and breaks the flow.

The only thing that doesn't work is this word extraction regex...

> 2. On the other hand, if you already have lots of `$->' occurrences in
> your text and you want to change them to right arrows, consider using
> `query-replace' or `replace-string' instead.

That is not what I want.

Re: Using punctuation in abbrev

[Emanuel Berg]

Aurélien Aptel <aurelien.aptel+emacs@gmail.com> writes:

> I want to substitute $-> with a unicode rightwards arrow (U+2192
> →).

I tried to do something similar, namely to replace "..." by
"&hellip;", and "---" by "&mdash;", in HTML-mode. It didn't work,
so I posted a question:

http://unix.stackexchange.com/q/71101

The reply was, this has to do with abbrev following the Emacs
definition of what is a word. Apparently, you can fiddle with
that, which, for a local mode, perhaps is smart (in my case, the
HTML mode), but I decided not to do it, because it might screw up
my cursor movement finger habits, and I am very picky about those.

-- 
Emanuel Berg - programmer (hire me! CV below)
computer projects: http://user.it.uu.se/~embe8573
internet activity: http://home.student.uu.se/embe8573

Re: Using punctuation in abbrev

[Emanuel Berg]

Aurélien Aptel <aurelien.aptel+emacs@gmail.com> writes:

> I plan on having a lot of easy-to-remember abbreviations for
> unicode chars I regularly use, all prefixed by $ or some other
> less used character ... Calling insert-char is too cumbersome
> and breaks the flow.

Just curious: what Unicode chars do you regularly use?

Another thought: Pick a shortcut that is very "fast" (short, and
without you having to move your hand), like M-; or C-o (it depends
what is already wired into your hands, and what you already do
useful stuff with - don't change that - but there are many
shortcuts...).

Then, write a defun that waits for a keystroke, then inserts the
Unicode char depending on what was hit, So, say, M-; l is right
arrow, M-; j is left arrow, M-; i is up arrow, etc.

-- 
Emanuel Berg - programmer (hire me! CV below)
computer projects: http://user.it.uu.se/~embe8573
internet activity: http://home.student.uu.se/embe8573

Re: Using punctuation in abbrev

[Stefan Monnier]

> Here's what I put in my init file:

>   (define-abbrev-table 'global-abbrev-table '(
>                                               ("$->" "→")
>                                               ("$=>" "⇒")
>                                               ("$foo" "X")
>                                               )
>     "custom abbrev table"
>     :regexp (rx (or bos bol space) (group (+ (not space)))))
>   (setq save-abbrevs nil)
>   (setq-default abbrev-mode t)

> But only the $foo substitution works and I can't figure out why.

Actually, if you call M-x expand-abbrev RET explicitly, you'll see they
all work.  What doesn't work for the first two is the "press space after
it to cause expansion", because this is only performed right after
a word char (hence it works for "$foo" because the last char is a word
char).


        Stefan

Re: Using punctuation in abbrev

[Aurélien Aptel]

On Sun, Jun 2, 2013 at 4:24 PM, Stefan Monnier <monnier@iro.umontreal.ca> wrote:
> Actually, if you call M-x expand-abbrev RET explicitly, you'll see they
> all work.  What doesn't work for the first two is the "press space after
> it to cause expansion", because this is only performed right after
> a word char (hence it works for "$foo" because the last char is a word
> char).

I see, thanks.
I've just looked at it: it's done in src/cmds.c:434 so behavior can't
be changed from Lisp, too bad.

Re: Using punctuation in abbrev

[Aurélien Aptel]

On Sun, Jun 2, 2013 at 7:21 AM, Emanuel Berg <embe8573@student.uu.se> wrote:
> Just curious: what Unicode chars do you regularly use?

mainly arrows, ×, ≠, ≈, ...

> Another thought: Pick a shortcut that is very "fast" (short, and
> without you having to move your hand), like M-; or C-o (it depends
> what is already wired into your hands, and what you already do
> useful stuff with - don't change that - but there are many
> shortcuts...).
>
> Then, write a defun that waits for a keystroke, then inserts the
> Unicode char depending on what was hit, So, say, M-; l is right
> arrow, M-; j is left arrow, M-; i is up arrow, etc.

That's not a bad alternative. I think I'll do this.

Re: Using punctuation in abbrev

[Andreas Röhler]

Am 02.06.2013 19:00, schrieb Aurélien Aptel:
> On Sun, Jun 2, 2013 at 4:24 PM, Stefan Monnier <monnier@iro.umontreal.ca> wrote:
>> Actually, if you call M-x expand-abbrev RET explicitly, you'll see they
>> all work.  What doesn't work for the first two is the "press space after
>> it to cause expansion", because this is only performed right after
>> a word char (hence it works for "$foo" because the last char is a word
>> char).
>
> I see, thanks.
> I've just looked at it: it's done in src/cmds.c:434 so behavior can't
> be changed from Lisp, too bad.
>
>

AFAICS the restriction is in abbrev--before-point for example, which fetches the abbrev using (backward-word 1).

Don't see a reason why not permit any printable char being part of an abbrev. Replace backward-word by (skip-chars-backward "^ \t\r\n\f")
etc.

Andreas

Re: Using punctuation in abbrev

[Stefan Monnier]

>> Actually, if you call M-x expand-abbrev RET explicitly, you'll see they
>> all work.  What doesn't work for the first two is the "press space after
>> it to cause expansion", because this is only performed right after
>> a word char (hence it works for "$foo" because the last char is a word
>> char).
> I see, thanks.
> I've just looked at it: it's done in src/cmds.c:434 so behavior can't
> be changed from Lisp, too bad.

The code that calls expand-abbrev from self-insert-command is indeed in
C, so you can't change it from Elisp, but that code also runs
post-self-insert-hook, so you can add a function to that hook which will
call expand-abbrev for the missing case.


        Stefan

Re: Using punctuation in abbrev

[Jambunathan K]

Aurélien Aptel <aurelien.aptel+emacs@gmail.com> writes:

> On Sun, Jun 2, 2013 at 7:21 AM, Emanuel Berg <embe8573@student.uu.se> wrote:
>> Just curious: what Unicode chars do you regularly use?
>
> mainly arrows, ×, ≠, ≈, ...

You can try this

    1. C-x b *scratch*
    2. M-x set-input-method RET TeX RET
    3. C-\  (M-x toggle-input-method RET)
    4. Type whatever in English
    5. C-\ (M-x toggle-input-method RET)
    6. \times \ne \approx
    7. C-\
    8. Type whatever in English

Step 6 should produce unicode chars for you.  They are easier to
remember. While in TeX input method, on a unicode "not equals" glyph if
you do,

   C-u C-x =

you will see this.  Just focus on "to input" field to get the TeX
encoding for that unicode character.

,----
|              position: 64 of 64 (98%), column: 0
|             character: ≠ (displayed as ≠) (codepoint 8800, #o21140, #x2260)
|     preferred charset: unicode (Unicode (ISO10646))
| code point in charset: 0x2260
|                script: symbol
|                syntax: . 	which means: punctuation
|              category: .:Base, c:Chinese, h:Korean, j:Japanese
|              to input: type "\neq" or "\ne" with TeX input method   
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
|           buffer code: #xE2 #x89 #xA0
|             file code: #xE2 #x89 #xA0 (encoded by coding system utf-8)
|               display: by this font (glyph code)
|     xft:-unknown-DejaVu Sans Mono-normal-normal-normal-*-20-*-*-*-m-0-iso10646-1 (#x7F6)
| 
| Character code properties: customize what to show
|   name: NOT EQUAL TO
|   general-category: Sm (Symbol, Math)
|   decomposition: (61 824) ('=' '̸')
| 
| There are text properties here:
|   fontified            t
`----

RE: Using punctuation in abbrev

[Drew Adams]

> >> I want to substitute $-> with a unicode rightwards arrow (U+2192  →).
> >
> > 1. Does that mean that you want to type `$->' and use `expand-abbrev'
> > each time to insert a Unicode right arrow?  If so, and if you don't
> > really care whether you use abbrev, consider binding insertion of a
> > right arrow char to a key sequence instead.
> 
> The abbrev minor mode is active in all my buffers and it doesn't have
> to be triggered by a specific key, it's automatically substituted as I
> type. I plan on having a lot of easy-to-remember abbreviations for
> unicode chars I regularly use, all prefixed by $ or some other less
> used character so that it only triggers when I want. Calling
> insert-char is too cumbersome and breaks the flow.

You apparently really want to use abbrev here.  OK.  But to be clear,
I did not suggest that you call `insert-char' each time.  Not at all.

I suggested that you bind a specific right-arrow inserting command
to a key, and hit that key each time, instead of typing `$->' each time.
IOW, in effect, add a Unicode right-arrow key to your keyboard.

Re: Using punctuation in abbrev

[Emanuel Berg]

Aurélien Aptel <aurelien.aptel+emacs@gmail.com> writes:

>> Just curious: what Unicode chars do you regularly use?
>
> mainly arrows, ×, ≠, ≈, ...

Aha, so you are writing scientific stuff, or perhaps documentation
with flow charts etc.? Or is there actually some new super
programming language which take advantage of all that notation?
(That would be... scary and impressive, at the same time.)

>> Another thought: Pick a shortcut that is very "fast" (short,
>> and without you having to move your hand), like M-; or C-o (it
>> depends what is already wired into your hands, and what you
>> already do useful stuff with - don't change that - but there
>> are many shortcuts...).  Then, write a defun that waits for a
>> keystroke, then inserts the Unicode char depending on what was
>> hit, So, say, M-; l is right arrow, M-; j is left arrow, M-; i
>> is up arrow, etc.
>
> That's not a bad alternative. I think I'll do this.

I'm happy you think so. Whatever solution you come up with, be
sure to post it here so all can learn from it :)

-- 
Emanuel Berg - programmer (hire me! CV below)
computer projects: http://user.it.uu.se/~embe8573
internet activity: http://home.student.uu.se/embe8573

Re: Using punctuation in abbrev

[Stefan Monnier]

> with flow charts etc.? Or is there actually some new super
> programming language which take advantage of all that notation?
> (That would be... scary and impressive, at the same time.)

If you're into that kind of thing, you might like Agda (note that it's
a proof assistant, so not useful if you want to do Octave/Mathematica
kind of things).


        Stefan

Re: Using punctuation in abbrev

[Aurélien Aptel]

On Sun, Jun 2, 2013 at 9:48 PM, Emanuel Berg <embe8573@student.uu.se> wrote:
> Aurélien Aptel <aurelien.aptel+emacs@gmail.com> writes:
>
>>> Just curious: what Unicode chars do you regularly use?
>>
>> mainly arrows, ×, ≠, ≈, ...
>
> Aha, so you are writing scientific stuff, or perhaps documentation
> with flow charts etc.? Or is there actually some new super
> programming language which take advantage of all that notation?
> (That would be... scary and impressive, at the same time.)

Scientific or programming related stuff yes. And about the programming
language, it has already been done 60 years ago with APL [1] which is
actually scary and impressive :)

1: https://en.wikipedia.org/wiki/APL_%28programming_language%29

Re: Using punctuation in abbrev

[Emanuel Berg]

Aurélien Aptel <aurelien.aptel+emacs@gmail.com> writes:

> Scientific or programming related stuff yes. And about the
> programming language, it has already been done 60 years ago with
> APL [1] which is actually scary and impressive :)

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> If you're into that kind of thing, you might like Agda (note
> that it's a proof assistant, so not useful if you want to do
> Octave/Mathematica kind of things).

Thanks for cool facts, but no, I'm not into that at all. I have a
degree in CS so I'm done with math and formal stuff (with the
exception of teaching it to school kids to get food money) - I
want to do Lisp and C, steady, everyday stuff. (And I don't
consider Lisp "functional" but rather "do whatever you want".) But
I can se some perks with such a "Unicode language" - for example,
how many ways have you seen "not equal" been syntaxed in PLs? !=,
<>, ne, =/= (Erlang), ... - what a mess, and one of the most
annoying things to have to stop and look up. Unicode notation
would help in such cases, for sure. And, the code would be more
pleasant to read (and more readable to non-programmers with a
background in math and/or natural sciences).

-- 
Emanuel Berg - programmer (hire me! CV below)
computer projects: http://user.it.uu.se/~embe8573
internet activity: http://home.student.uu.se/embe8573

Re: Using punctuation in abbrev

[Yuri Khan]

On Mon, Jun 3, 2013 at 12:17 AM, Aurélien Aptel
<aurelien.aptel+emacs@gmail.com> wrote:
> On Sun, Jun 2, 2013 at 7:21 AM, Emanuel Berg <embe8573@student.uu.se> wrote:
>> Just curious: what Unicode chars do you regularly use?
>
> mainly arrows, ×, ≠, ≈, ...

This is an Emacs list and you are asking how to enter characters in
Emacs, but I feel this is better solved at the system level (and I
actually believe there is a system outside Emacs) so that you can
enter these characters in other programs.

* If you are running Emacs on X11, you can hack your XKB configuration
to put frequently-needed characters on Level3 (which is an alternate
shift state often configured to activate on right Alt or another
modifier key of your choice). You might even find an existing symbols
map that has characters you need.

* On Windows, the equivalent is the MSKLC tool.

* I believe there is an analogous tool on Mac, too.

Re: Using punctuation in abbrev

[Emanuel Berg]

Yuri Khan <yuri.v.khan@gmail.com> writes:

> * If you are running Emacs on X11, you can hack your XKB
> configuration to put frequently-needed characters on Level3
> (which is an alternate shift state often configured to activate
> on right Alt or another modifier key of your choice). You might
> even find an existing symbols map that has characters you need.

If you run Emacs in a Linux VT, you can do approximately the same
with loadkeys (read the part on strings in the man page).

However, I don't see why this should be done outside Emacs, if
Emacs is the place, where the OP wants to use it?

By the way - you have a cool name :)

-- 
Emanuel Berg - programmer (hire me! CV below)
computer projects: http://user.it.uu.se/~embe8573
internet activity: http://home.student.uu.se/embe8573

Re: Using punctuation in abbrev

[Yuri Khan]

On Tue, Jun 4, 2013 at 1:05 AM, Emanuel Berg <embe8573@student.uu.se> wrote:

>> * If you are running Emacs on X11, you can hack your XKB
>> configuration […]
>
> […]I don't see why this should be done outside Emacs, if
> Emacs is the place, where the OP wants to use it?

Because if I were OP I would want to have access to special characters
all over the system, not just Emacs.

Re: Using punctuation in abbrev

[Aurélien Aptel]

On Tue, Jun 4, 2013 at 5:05 AM, Yuri Khan <yuri.v.khan@gmail.com> wrote:
> Because if I were OP I would want to have access to special characters
> all over the system, not just Emacs.
>

It's a good idea but x, => or -> is more natural and easier to remember for me.

Re: Using punctuation in abbrev

[Emanuel Berg]

Yuri Khan <yuri.v.khan@gmail.com> writes:

>>> * If you are running Emacs on X11, you can hack your XKB
>>> configuration […]
>>
>> ... I don't see why this should be done outside Emacs, if Emacs
>> is the place, where the OP wants to use it?
>
> Because if I were OP I would want to have access to special
> characters all over the system, not just Emacs.

Why? There is no reason to do programming and/or tech writing
outside Emacs, if Emacs is the place you normally do it. Also,
Emacs - with Elisp - gives you the tools to setup the execution of
such tasks.

The shell CLI should be used for the batch tools - sed, units,
those things - all typing, and editing, should be done in the
editor.

-- 
Emanuel Berg - programmer (hire me! CV below)
computer projects: http://user.it.uu.se/~embe8573
internet activity: http://home.student.uu.se/embe8573

Re: Using punctuation in abbrev

[Stefan Monnier]

> However, I don't see why this should be done outside Emacs, if
> Emacs is the place, where the OP wants to use it?

To each his own, but FWIW, I do prefer input methods that
work everywhere.  For this reason I don't use the latin-N input methods
and prefer to use "compose + ' + e" to enter é.

But it's often easier to add such special-purpose tricks to Emacs,
especially if you want them to be custom-tailored to the specific mode
in which you use it.

And in the case of the OP, since he wants it to work in all Emacs modes,
it's likely he'd also want it to work outside Emacs.


        Stefan

Re: Using punctuation in abbrev

[Emanuel Berg]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> However, I don't see why this should be done outside Emacs, if
>> Emacs is the place, where the OP wants to use it?
>
> To each his own

Yes. (Those are my opinions. I'm not telling anyone what to do.)

> but FWIW, I do prefer input methods that work everywhere.

Yes. I *prefer* that as well. In much the same way as I prefer the
dired colors to match the ls dircolors. The colors in the Linux VT
and the colors in urxvt. The cursor movements in those terminals,
and in Emacs. And so on, forever (almost).

But, because I've spent so much time on such "unification" issues,
I have come to the conclusion that sometimes, although preferable,
it is not worth a huge effort just for the sake of it. And I think
this is such a case.

> For this reason I don't use the latin-N input methods and prefer
> to use "compose + ' + e" to enter é.

Here, I agree without reservation. I use the US layout and for the
special chars of my native language I, as well, rely on the
compose key. You don't have to type that much to "learn anew". It
is like people loosing one hand in an accident. Say they are right
handed, and they lose that hand. Before long, they are as good
with their lefts, as they ever were with their
rights. Adaptability - perhaps the most human, of all humans
things...

> But it's often easier to add such special-purpose tricks to
> Emacs, especially if you want them to be custom-tailored to the
> specific mode in which you use it.

Yes, I said this to (almost). A lot of stuff that you do almost
routinely in Emacs and Elisp - when faced with the same problem,
say, at shell level, I am always surprised how difficult
everything is. Scripts, functions, the PATH, where to put
documentation - I don't know if I am just worse at it, or if Emacs
is actually better (if that could be quantified somehow). Anyway,
that's always the case for me.

-- 
Emanuel Berg - programmer (hire me! CV below)
computer projects: http://user.it.uu.se/~embe8573
internet activity: http://home.student.uu.se/embe8573

Re: Using punctuation in abbrev

[Yuri Khan]

On Wed, Jun 5, 2013 at 12:27 AM, Emanuel Berg <embe8573@student.uu.se> wrote:
> Yuri Khan <yuri.v.khan@gmail.com> writes:
>> Because if I were OP I would want to have access to special
>> characters all over the system, not just Emacs.
>
> Why? There is no reason to do programming and/or tech writing
> outside Emacs, if Emacs is the place you normally do it. Also,
> Emacs - with Elisp - gives you the tools to setup the execution of
> such tasks.

Because such things may need to be discussed with other people, and
this involves an email client, a web browser, an instant messenger, or
any combination of the above. Yes, I’m aware Emacs has solutions for
each of those, and no, I do not want to give up Firefox, Thunderbird
and Gajim. That’s what I meant when I said I believe in an operating
system/environment outside Emacs.

Re: Using punctuation in abbrev

[Emanuel Berg]

Yuri Khan <yuri.v.khan@gmail.com> writes:

>>> Because if I were OP I would want to have access to special
>>> characters all over the system, not just Emacs.
>>
>> Why? There is no reason to do programming and/or tech writing
>> outside Emacs, if Emacs is the place you normally do it. Also,
>> Emacs - with Elisp - gives you the tools to setup the execution
>> of such tasks.
>
> Because such things may need to be discussed with other people,
> and this involves an email client, a web browser, an instant
> messenger, or any combination of the above. Yes, I’m aware Emacs
> has solutions for each of those, and no, I do not want to give
> up Firefox, Thunderbird and Gajim. That’s what I meant when I
> said I believe in an operating system/environment outside Emacs.

Well, if you use all those applications, a lot of things will be
different in a lot of ways. If you have no problem with that, you
might say that input is unique in the sense that with everything
else not so, at least input should be consistent, always and
everywhere. In that sense, I agree.

-- 
Emanuel Berg - programmer (hire me! CV below)
computer projects: http://user.it.uu.se/~embe8573
internet activity: http://home.student.uu.se/embe8573