.emacs poser

.emacs poser

[B. T. Raven]

In my emacs I have these global-set-key forms:

....

(global-set-key "\C-cg" (lambda () (interactive) (insert  ?° )))
(global-set-key "\C-ch" (lambda () (interactive)  (insert  ?·)))
(global-set-key "\C-c-" (lambda () (interactive)  (insert  ?— )))
(global-set-key "\C-cI" (lambda () (interactive)  (insert  ?‽ ))) ;; ;;
;; capital eye interrobang
(global-set-key "\C-cL" (lambda () (interactive)  (insert  ?£ )))
(global-set-key "\C-cm" (lambda () (interactive)  (insert  ?ˉ)))
;;(global-set-key "\C-cM" (lambda () (interactive)  (insert  ?̄✠)))
(global-set-key "\C-cp" (lambda () (interactive) (insert  ?¶ )))
(global-set-key "\C-cr" (lambda () (interactive)  (insert  ?® )))
(global-set-key "\C-cs" (lambda () (interactive) (insert  ?§ )))
(global-set-key "\C-ci" (lambda() (interactive) (insert
(format-time-string "%a %Y .....

These and many others work fine (in w32 native v. 23.3) but any attempt
to evaluate "\C-cM" (capital M) truncates the rest of the .emacs and
wakes the Debugger:

Debugger entered--Lisp error: (invalid-read-syntax "?")
  read(#<buffer .emacs>)
  preceding-sexp()
  eval-last-sexp-1(nil)
  eval-last-sexp(nil)
  call-interactively(eval-last-sexp nil nil)
  recursive-edit()
  byte-code("\306\x10	@\307=\203! ......

the glyph shown is the Maltese cross but any character after the ? fails
to work in the same way. Other capitals are distinguished, just not
upper case M.

Does any of you have any idea whats going on here. With just that line
commented out the .emacs loads with no problem.

Thanks,

Ed

Re: .emacs poser

[Emanuel Berg]

"B. T. Raven" <btraven@nihilo.net> writes:

> ...

I don't think the keystroke has anything to do with
it. It is the insert command that follows, or more
specifically that char, whatever it is, that breaks the
file.

Test this:

(local-set-key "\C-cM"
  '(lambda ()
   (interactive)
   (insert "Big M is not the problem") ))

Why do you need those goofy chars to begin with?

And - it is a good idea to make the subject header more
informative.

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

Re: .emacs poser

[Dale Snell]

On Mon, 16 Dec 2013 18:01:02 -0600
"B. T. Raven" <btraven@nihilo.net> wrote:

> In my emacs I have these global-set-key forms:
> 
> ....
> 
> (global-set-key "\C-cg" (lambda () (interactive) (insert  ?° )))
> (global-set-key "\C-ch" (lambda () (interactive)  (insert  ?·)))
> (global-set-key "\C-c-" (lambda () (interactive)  (insert  ?— )))
> (global-set-key "\C-cI" (lambda () (interactive)  (insert  ?‽
> ))) ;; ;; ;; capital eye interrobang
> (global-set-key "\C-cL" (lambda () (interactive)  (insert  ?£ )))
> (global-set-key "\C-cm" (lambda () (interactive)  (insert  ?ˉ)))
> ;;(global-set-key "\C-cM" (lambda () (interactive)  (insert  ?̄✠)))
> (global-set-key "\C-cp" (lambda () (interactive) (insert  ?¶ )))
> (global-set-key "\C-cr" (lambda () (interactive)  (insert  ?® )))
> (global-set-key "\C-cs" (lambda () (interactive) (insert  ?§ )))
> (global-set-key "\C-ci" (lambda() (interactive) (insert
> (format-time-string "%a %Y .....
> 
> These and many others work fine (in w32 native v. 23.3) but any
> attempt to evaluate "\C-cM" (capital M) truncates the rest of
> the .emacs and wakes the Debugger:

==========%<---------- SNIP! ---------->%========== 

I don't know why this isn't working for you, I'm an E-Lisp tyro.
Still, it looks to me like you're re-inventing the wheel.  Rather than
pulling your hair out over this, try using iso-transl.elc (comes with
Emacs) or perhaps typopunct.el (not part of Emacs proper).  Check
EmacsWiki for information on both of them.  Both can be extended to add
such characters as Registered ® and the Maltese Cross ✠.  ("C-x 8 * R"
and "C-x 8 * M", respectively, using iso-transl.  I added both
those characters myself.  It's very easy.)

Hope this helps.

--Dale

--
"Text processing has made it possible to right-justify any idea,
even one which cannot be justified on any other grounds."
		-- J. Finnegan, USC.

Re: .emacs poser

[Emanuel Berg]

Dale Snell <ddsnell@frontier.com> writes:

> Registered ® and the Maltese Cross ✠.  ("C-x 8 * R"
> and "C-x 8 * M", respectively, using iso-transl.  I
> added both those characters myself.  It's very easy.)

It might be easy to *add* such keystrokes, but they are
not easy to hit. We are typists, not accordion players.

Again, why would anyone need those chars on a regular
basis? (Not a rhetorical question!)

If you, in certain *words* need those chars, how about
using abbrev to do that for you? So you type the word
without the special char(s), and abbrev fixes it as you
hit space (or some other delimiter): e.g., you type
Marquez to get Márquez.

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

RE: .emacs poser

[Drew Adams]

> In my emacs I have these global-set-key forms:
> (global-set-key "\C-cg" (lambda () (interactive) (insert  ?° )))
> (global-set-key "\C-ch" (lambda () (interactive)  (insert  ?·)))

An easy way to define commands that insert individual Unicode chars
that you use often is to use library `ucs-cmds.el'.

* Command `ucsc-insert' does the same as `insert-char': inserts a
  char you choose by its Unicode name (or hex code) using completion.
  But with a negative prefix arg it also defines a command with the
  same name.

* Macro `ucs-make-commands' defines a whole set of such commands,
  based on a regexp that is matched against all Unicode char names.

You can then bind the commands to keys.  E.g.:

  C-- M-x ucsc-insert greek small letter lambda RET

defines command `greek-small-letter-lambda', which inserts that
character (or N such chars, where N is the prefix arg).

Then: (global-set-key "\C-cl 'greek-small-letter-lambda).

http://www.emacswiki.org/emacs-en/download/ucs-cmds.el

http://www.emacswiki.org/emacs/UnicodeEncoding#ucsc-insert

Re: .emacs poser

[Dale Snell]

On Tue, 17 Dec 2013 03:03:26 +0100
Emanuel Berg <embe8573@student.uu.se> wrote:

> Dale Snell <ddsnell@frontier.com> writes:
> 
> > Registered ® and the Maltese Cross ✠.  ("C-x 8 * R"
> > and "C-x 8 * M", respectively, using iso-transl.  I
> > added both those characters myself.  It's very easy.)
> 
> It might be easy to *add* such keystrokes, but they are
> not easy to hit. We are typists, not accordion players.

Once upon a time, in my mis-spent youth, I was.  :-)
That said, the only chord there is the C-x.  The rest are a
sequence of keystrokes.  Err, well, on a US keyboard anyway.  I
can't speak for European keyboards.

> Again, why would anyone need those chars on a regular
> basis? (Not a rhetorical question!)

It depends on what your use-case is.  Personally, I type up a lot
of recipes, which usually want a temperature setting.  Hence I use
the degree symbol (°) a lot.  I have occasional use for super- and
sub-script characters, and the ever popular copyright (©),
trademark (™), and registered trademark signs (®).  If I need
anything more demanding, like en and em dashes, or primes instead
of quotes, I'll fire up a text processor and go that way.  Of
course, in a purely text forum, like this, someone would
undoubtedly complain if I were to attach the entire message as a
PDF.

There are certain organizations that want their documents written
in a certain format, which may include Pilcrow and Section marks,
and other such things.  Happily, I don't deal with those.  (Again,
I'd use LaTeX or Groff for that.)

> If you, in certain *words* need those chars, how about
> using abbrev to do that for you? So you type the word
> without the special char(s), and abbrev fixes it as you
> hit space (or some other delimiter): e.g., you type
> Marquez to get Márquez.

That's limiting yourself.  If you need accented characters, learn
how to enter them in a general way, not just specific words.  Then
you can use them anywhere.  As I mentioned in my previous post,
check EmacsWiki for information about that.

Anyway, these are all just my opinions.  If you prefer a different
way, more power to you.

--Dale

--
"The Enrichment Center reminds you that the Weighted Companion Cube
will never threaten to stab you and, in fact, cannot speak.  In the
event that the Weighted Companion Cube does speak, the Enrichment
Center urges you to disregard its advice."

Re: .emacs poser

[Yuri Khan]

On Tue, Dec 17, 2013 at 7:01 AM, B. T. Raven <btraven@nihilo.net> wrote:

> In my emacs I have these global-set-key forms:

IMO you’re doing it wrong, by solving the general problem of entering
special character for a specific case of Emacs.

> ;;(global-set-key "\C-cM" (lambda () (interactive)  (insert  ?̄✠)))

> the glyph shown is the Maltese cross but any character after the ? fails
> to work in the same way.

What I see[1] is Maltese Cross and a Macron over the preceding
character. That’s two code points, not one. You need to either remove
the macron, or treat the two characters as a string, enclosing it in
quotes instead of prefixing with a question mark.

[1]: I actually see the macron over the Maltese cross, but that’s due
to a bug in font rendering on my system. Diacritic marks are supposed
to affect the preceding character, and in your message the macron code
point appears between the question mark and the Maltese cross.

Re: .emacs poser

[Emanuel Berg]

Dale Snell <ddsnell@frontier.com> writes:

> That said, the only chord there is the C-x.  The rest
> are a sequence of keystrokes.

It doesn't matter. It is too many hits to do too
little.

> I can't speak for European keyboards.

The US layout is better for programming but those chars
are not as goofy as those you use. Semi-colon, and all
the brackets, are better placed on the US layout
keyboard, and for whatever language-specific chars you
need, there is the compose key.

Which by the way is another solution that I think is
much better than setting this up in Emacs.

> and the ever popular copyright (©), trademark (™),
> and registered trademark signs (®).

Serious?

> If I need anything more demanding, like en and em
> dashes, or primes instead of quotes, I'll fire up a
> text processor

A word processor? Like OpenOffice or Word?

> and go that way.  Of course, in a purely text forum,
> like this, someone would undoubtedly complain if I
> were to attach the entire message as a PDF.

Yes. PDF is for things to be *printed*, like manuals,
or scientific work with need for special notation.

> There are certain organizations that want their
> documents written in a certain format, which may
> include Pilcrow and Section marks, and other such
> things.  Happily, I don't deal with those.  (Again,
> I'd use LaTeX or Groff for that.)

Groff! Wow, you are a man (pun) of many surprises. Is
that used outside of the Unix manpages world?

LaTeX is great obviously. I would drop the word
processor and use Emacs (or Vim) + LaTeX.

> That's limiting yourself.  If you need accented
> characters, learn how to enter them in a general way,
> not just specific words.

It is not about *ability*, it is about *speed* and
*ergonomics* and *limiting the mental effort* when
doing a routine thing, as typing. To memorize and type
some four or five hit combination just to get a goofy
char that is (almost) never used doesn't make sense.

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

Re: .emacs poser

[Dale Snell]

On Tue, 17 Dec 2013 04:16:04 +0100
Emanuel Berg <embe8573@student.uu.se> wrote:

> Dale Snell <ddsnell@frontier.com> writes:
> 
> > I can't speak for European keyboards.
> 
> The US layout is better for programming but those chars
> are not as goofy as those you use. Semi-colon, and all
> the brackets, are better placed on the US layout
> keyboard, and for whatever language-specific chars you
> need, there is the compose key.

The compose key is good.  Now if I could just set my system up so
that the Windows-Menu key worked as a compose key in both X and
Linux VTs.

> Which by the way is another solution that I think is
> much better than setting this up in Emacs.
> 
> > and the ever popular copyright (©), trademark (™),
> > and registered trademark signs (®).
> 
> Serious?

Serious, unfortunately.  Twas Corporate Policy.  Any brand or
trade name must be given the proper attribution and trademark
symbol.  So things like "Microsoft Word" became "Microsoft®
Word®".  Every. Freaking. Time.  Arrgh.

> > If I need anything more demanding, like en and em
> > dashes, or primes instead of quotes, I'll fire up a
> > text processor
> 
> A word processor? Like OpenOffice or Word?

Gak!  No!  Mother Hitton's Littul Kittons, no!  TEXT processor.  I
loathe word processores, Word, LibreOffice, Abiword, whatever.
They give me the screaming heebie-geebies whenever I have to use
one of them.  Give me Emacs and Groff or TeX.

> > There are certain organizations that want their
> > documents written in a certain format, which may
> > include Pilcrow and Section marks, and other such
> > things.  Happily, I don't deal with those.  (Again,
> > I'd use LaTeX or Groff for that.)
> 
> Groff! Wow, you are a man (pun) of many surprises. Is
> that used outside of the Unix manpages world?

Well, obviously I use it.  (I find Groff with the Mom macro set
very comfortable to use when writing prose.)  But yes, there are
folks who don't write man pages who use Groff.  Some write books
with it, both technical and fiction.  It's not as flexible as TeX,
but then, what is?  That said, if I had to write a math-heavy
book, or one with lots of nuclear chemistry, I'd use TeX.  Groff
and eqn are nice for light math work, but TeX is the Queen! of
beautiful math.

> LaTeX is great obviously. I would drop the word
> processor and use Emacs (or Vim) + LaTeX.

If I didn't have Groff, I'd go with LaTeX, definitely.  With
Emacs, of course.  (BTW, does anybody still use Lout?)

> > That's limiting yourself.  If you need accented
> > characters, learn how to enter them in a general way,
> > not just specific words.
> 
> It is not about *ability*, it is about *speed* and
> *ergonomics* and *limiting the mental effort* when
> doing a routine thing, as typing. To memorize and type
> some four or five hit combination just to get a goofy
> char that is (almost) never used doesn't make sense.

Time for the compose key?  :-)  Honestly, I've used the "C-x 8"
sequence so much it's second nature by now.

But I think we'll have to agree to disagree, here.  I am going to
investigate Drew's ucs-cmds.el library; it looks interesting.
Besides, I think we're getting a bit off-topic here.

--Dale

--
I recall hearing that highly-classified data must be destroyed by
physically shredding the medium.  Yes, throw your disk drive in
the shredder!  (Just imagine the class of machinery required to
digest an RA81 HDA.)  -- Mark Wood on linux-kernel

Re: .emacs poser

[Jambunathan K]

I would recommend Org-mode with export to LaTeX and PDF.  You can create
OpenDocument text or create HTML for posting to a blog.

Dale Snell <ddsnell@frontier.com> writes:

> Hence I use the degree symbol (°) a lot.  I have occasional use for
> super- and sub-script characters, and the ever popular copyright (©),
> trademark (™), and registered trademark signs (®).

You can do a

   M-x set-input-method TeX

then typing the string on the first column will give you the characters
on the second column

    | what to type | what you get |
    |--------------+--------------|
    | \degree      | °            |
    | \copyright   | ©            |
    | ^{TM}        | ™            |
    | \registered  | ®            |
    | ---          | —            |
    | --           | –            |


To check what to type, do

   C-u C-x =

on a character inserted with C-x 8 C-m and look for the following line
in the *Help* buffer

  to input: type "--" with TeX input method

Ofcourse you can conjure up your own abbrevs and not go with above
system provided defaults.

   (info "(emacs) Defining Abbrevs")

Re: .emacs poser

[Tilman Ahr]

Emanuel Berg <embe8573@student.uu.se> writes:

> Dale Snell <ddsnell@frontier.com> writes:

[...]

>> I can't speak for European keyboards.
>
> The US layout is better for programming
[...]

Arguably, but for most layouts and programming languages, that probably
hosds true.

> Semi-colon, and all the brackets, are better placed on the US layout
> keyboard,

I'm not convinced of that, at least for the DE-DE keyboard layout.

> and for whatever language-specific chars you need, there is the
> compose key.

Although I'd *hate* to have to use that to get Ä,Ö,Ü and ß. I use those
*a lot* when writing in German, and the DE-DE layout makes them
accessible by a single Keystroke.

I do have the „dead keys“ option enabled (to make typing the usual
accents for french, at least, feasible without much ado) and remap
CAPSLOCK[1] to the compose key, nonetheless. 

> Which by the way is another solution that I think is
> much better than setting this up in Emacs.

Yes. Definitely. But there are (coff, coff) Operating Systems that make
using a compose key hard, if not impossible… 

>> and the ever popular copyright (©), trademark (™),
>> and registered trademark signs (®).
>
> Serious?

Depending on the area one works in, and considering how the US civil
justice system works(fsvo), I could see that. Personally, I know how to
create © and ® on my keyboard, how I'd get the [TM] symbol, I have no
Idea. I believe I recall it's trivial with a compose key, and should be
doable without, but I don't need that symbol other than for ironic
remarks on usenet, and there, [TM] does the trick quite satisfactorily.

>> If I need anything more demanding, like en and em
>> dashes,

Those, I use frequently. And real quote marks as demanded by German
typographic conventions. I don't even need the compose key for
them. Other than the en-dash…

>> or primes instead of quotes, I'll fire up a text processor
>
> A word processor? Like OpenOffice or Word?

I'd guess he actually meant a text processor, like TeX/LaTeX

>> and go that way.  Of course, in a purely text forum,
>> like this, someone would undoubtedly complain if I
>> were to attach the entire message as a PDF.
>
> Yes. PDF is for things to be *printed*, like manuals,
> or scientific work with need for special notation.

Actually, PDF allows for very neat features used for
on-screen-presentation and even has some rather sophisticated features
for dealing with textual matter. Those are just aspects of that format
that never took off.

So, in everyday use: Yes, it's mainly used as a format for well-defined
visual representation of … data. And in this day and age of the
paperless office, these are usually printed out.

I had a job interview last week where the interviewer had my e-mail with
all the relevant stuff for my application (CV, Certificates, References
and all that) as a PDF on his PCs screen right in front of him, but
still printed it out. On good paper, using a colour ink-jet printer that
went to photo mode, because there was a photo on the CV (most german
employers really appreciate a CV with a decent photo. No idea why). That
whole printout went straight to the bin fifteen minutes later, btw.

>> There are certain organizations that want their
>> documents written in a certain format, which may
>> include Pilcrow and Section marks, and other such
>> things.  Happily, I don't deal with those.  (Again,
>> I'd use LaTeX or Groff for that.)
>
> Groff! Wow, you are a man (pun) of many surprises. Is
> that used outside of the Unix manpages world?

It used to be. And I remember coming across a couple of projects where
it was used to nicely format simple textual data in an automated way
even a few years ago…

> LaTeX is great obviously. I would drop the word
> processor and use Emacs (or Vim) + LaTeX.

Wellyes. No. For a one-off, one to ten pages, needs to look somewhat
presentable thingie, I'll take LO, thank you. If it actually needs to
look *good*, or be reusable for different documents with a similar
layout (visually orientated), I'll use Scribus, or whatever other
DTP-Software I can get my hands on. (Restaurant menues would be one
example that comes to mind)

If it's about presenting a large(ish) amount of text in a well-typeset,
neat way (Books and the like), or about automating stuff (like
generating invoices from a database), TeX/LaTeX rules supreme.

>> That's limiting yourself.  If you need accented
>> characters, learn how to enter them in a general way,
>> not just specific words.
>
> It is not about *ability*, it is about *speed* and
> *ergonomics* and *limiting the mental effort* when
> doing a routine thing, as typing.

Yes. I like abbrev-mode for larger projects. Takes a while to get that
up to speed, but when it's ready, it flies.

> To memorize and type some four or five hit combination just to get a
> goofy char that is (almost) never used doesn't make sense.

That combination of a non-US layout with dead keys makes most of what I
use regularly a breeze. For really rare stuff, Compose works well. It's
not really that hard to remember the more usual combinations.

Footnotes: 
[1]  I have no idea why anybody would have a use for that key's original
     function on anything but a mechanic typewriter. Other than SHOUTING
     YOUR LUNGS OUT…

Re: .emacs poser

[Kevin Rodgers]

On 12/16/13 5:01 PM, B. T. Raven wrote:
> In my emacs I have these global-set-key forms:
>
> ....
>
> (global-set-key "\C-cg" (lambda () (interactive) (insert  ?° )))
> (global-set-key "\C-ch" (lambda () (interactive)  (insert  ?·)))
> (global-set-key "\C-c-" (lambda () (interactive)  (insert  ?— )))
> (global-set-key "\C-cI" (lambda () (interactive)  (insert  ?‽ ))) ;; ;;
> ;; capital eye interrobang
> (global-set-key "\C-cL" (lambda () (interactive)  (insert  ?£ )))
> (global-set-key "\C-cm" (lambda () (interactive)  (insert  ?ˉ)))
> ;;(global-set-key "\C-cM" (lambda () (interactive)  (insert  ?̄✠)))
> (global-set-key "\C-cp" (lambda () (interactive) (insert  ?¶ )))
> (global-set-key "\C-cr" (lambda () (interactive)  (insert  ?® )))
> (global-set-key "\C-cs" (lambda () (interactive) (insert  ?§ )))
> (global-set-key "\C-ci" (lambda() (interactive) (insert
> (format-time-string "%a %Y .....
>
> These and many others work fine (in w32 native v. 23.3) but any attempt
> to evaluate "\C-cM" (capital M) truncates the rest of the .emacs and
> wakes the Debugger:
>
> Debugger entered--Lisp error: (invalid-read-syntax "?")
>    read(#<buffer .emacs>)
>    preceding-sexp()
>    eval-last-sexp-1(nil)
>    eval-last-sexp(nil)
>    call-interactively(eval-last-sexp nil nil)
>    recursive-edit()
>    byte-code("\306\x10	@\307=\203! ......
>
> the glyph shown is the Maltese cross but any character after the ? fails
> to work in the same way. Other capitals are distinguished, just not
> upper case M.
>
> Does any of you have any idea whats going on here. With just that line
> commented out the .emacs loads with no problem.

There is some other character between "?" and "✠" on that line.

BTW, instead of including the character literally, you could use a
representation that doesn't rely the file encoding (and leave the literal
character in a comment):

(global-set-key "\C-cM" (lambda () (interactive)  (insert ?\x2720))) ; U+2720 = ✠

or put the literal character in a doc string that you can access with
`C-h k' (M-x describe-key) or `C-h d' (M-x apropos-documentation):

(global-set-key "\C-cM" (lambda ()
			  "Insert U+2720 = \"✠\"."
			  (interactive)
			  (insert ?\x2720)))

-- 
Kevin Rodgers
Denver, Colorado, USA

RE: .emacs poser

[Drew Adams]

> > In my emacs I have these global-set-key forms:
> 
> IMO you’re doing it wrong, by solving the general problem of entering
> special character for a specific case of Emacs.

There is no wrong or right way here.  Does your keyboard have a key
that inserts the character `2' (yes)?  Is dedicating that key
sequence to inserting that one character wasteful or "doing it wrong"?

Binding a command that inserts a commonly used character to a key
can be quite handy, regardless of whether there is also a general
way to insert any character.  You *could* insert character `2'
using `C-x 8 RET digit two' each time, but why bother?

Think of binding such keys as one way to extend your keyboard, or
even to change it from one language to another, or...

RE: .emacs poser

[Doug Lewan]

I'm surprised that there's only one mention of input methods and so much about composing characters, which is, as you point out, a little clumsy.

It's easy to extend an input method with new rules. I frequently use the latin-1-prefix mode for a handful of special characters that I use. The lisp to add my rules looks like this:

(setq default-input-method "latin-1-prefix")
(let ((quail-setup-buf (get-buffer-create "*Quail setup*")))
  (save-excursion (set-buffer quail-setup-buf)
		  (toggle-input-method)
		  ;; MAINTENANCE Look for SYMBOLS
		  (quail-defrule ":A" ?∀)
		  (quail-defrule ":E" ?∃)
		  (quail-defrule ":e" ?∈)
                                     ...))

I also notice this comment near that code:
;; This seems rather a bit of a hack.
That was my sense long ago; I don't know if that's really a fair sentiment anymore.

I hope this helps.

,Douglas
Douglas Lewan
Shubert Ticketing
(201) 489-8600 ext 224

Statistics: The science of figuring out exactly how wrong you probably are.


-----Original Message-----
From: help-gnu-emacs-bounces+dougl=shubertticketing.com@gnu.org [mailto:help-gnu-emacs-bounces+dougl=shubertticketing.com@gnu.org] On Behalf Of B. T. Raven
Sent: Monday, 2013 December 16 19:01
To: help-gnu-emacs@gnu.org
Subject: .emacs poser

In my emacs I have these global-set-key forms:

....

(global-set-key "\C-cg" (lambda () (interactive) (insert  ?° )))
(global-set-key "\C-ch" (lambda () (interactive)  (insert  ?·)))
(global-set-key "\C-c-" (lambda () (interactive)  (insert  ?— )))
(global-set-key "\C-cI" (lambda () (interactive)  (insert  ?‽ ))) ;; ;;
;; capital eye interrobang
(global-set-key "\C-cL" (lambda () (interactive)  (insert  ?£ )))
(global-set-key "\C-cm" (lambda () (interactive)  (insert  ?ˉ)))
;;(global-set-key "\C-cM" (lambda () (interactive)  (insert  ?̄✠)))
(global-set-key "\C-cp" (lambda () (interactive) (insert  ?¶ )))
(global-set-key "\C-cr" (lambda () (interactive)  (insert  ?® )))
(global-set-key "\C-cs" (lambda () (interactive) (insert  ?§ )))
(global-set-key "\C-ci" (lambda() (interactive) (insert
(format-time-string "%a %Y .....

These and many others work fine (in w32 native v. 23.3) but any attempt
to evaluate "\C-cM" (capital M) truncates the rest of the .emacs and
wakes the Debugger:

Debugger entered--Lisp error: (invalid-read-syntax "?")
  read(#<buffer .emacs>)
  preceding-sexp()
  eval-last-sexp-1(nil)
  eval-last-sexp(nil)
  call-interactively(eval-last-sexp nil nil)
  recursive-edit()
  byte-code("\306\x10	@\307=\203! ......

the glyph shown is the Maltese cross but any character after the ? fails
to work in the same way. Other capitals are distinguished, just not
upper case M.

Does any of you have any idea whats going on here. With just that line
commented out the .emacs loads with no problem.

Thanks,

Ed

Re: .emacs poser

[Sebastien Vauban]

Kevin Rodgers wrote:
> [...] put the literal character in a doc string that you can access
> with `C-h k' (M-x describe-key) or `C-h d' (M-x
> apropos-documentation):
>
> (global-set-key "\C-cM" (lambda ()
> 			  "Insert U+2720 = \"✠\"."
> 			  (interactive)
> 			  (insert ?\x2720)))

I see that your using \xABCD notation while I do have Unicode characters
encoded as \uABCD in my .emacs file.

I've tried to find information about that in the Emacs manual but saw no
reference to either \x or \u notations.

What's the difference?

Best regards,
  Seb

-- 
Sebastien Vauban

Re: .emacs poser

[Emanuel Berg]

> The compose key is good.  Now if I could just set my
> system up so that the Windows-Menu key worked as a
> compose key in both X and Linux VTs.

Try this:

XKBOPTIONS=compose:ralt # change ralt to preferred key

in /etc/default/keyboard

then

sudo dpkg-reconfigure console-setup

then, in /etc/console-setup/remap.inc

### compose key
## current state: `dumpkeys --compose-only'
compose 'o' 'a' to U+00E5 # å
compose 'o' 'A' to U+00C5 # Å
compose '0' 'a' to U+00E5 # å
compose '0' 'A' to U+00C5 # Å
compose '"' 'a' to U+00E4 # ä
compose '"' 'A' to U+00C4 # Ä
compose '/' 'o' to U+00F3 # ó
compose '/' 'O' to U+00D3 # Ó
compose '"' 'o' to U+00F6 # ö
compose '"' 'O' to U+00D6 # Ö
compose '/' 'e' to U+00E9 # é
compose '/' 'E' to U+00C9 # É
compose '/' 'a' to U+00E1 # á
compose '/' 'A' to U+00C1 # Á
compose '/' 'u' to U+00FA # ú
compose '/' 'U' to U+00DA # Ú
compose '/' 'i' to U+00ED # í
compose '/' 'I' to U+00CD # Í
compose '\' 'a' to U+00E0 # à
compose '\' 'A' to U+00C0 # À

then

loadkeys -c -s /etc/console-setup/remap.inc # same file

The compose key works for me in X as well, though I
don't remember how I set that up, if indeed I did.

> Serious, unfortunately.  Twas Corporate Policy.  Any
> brand or trade name must be given the proper
> attribution and trademark symbol.  So things like
> "Microsoft Word" became "Microsoft® Word®".
> Every. Freaking. Time.  Arrgh.

This is a prime example when you should use an
abbrev. Call it msw, for example. You probably already
know how to setup an abbrev, but I mention it
nonetheless. Your example (MS Word) first, then some
other creative uses of abbrev.

(let*((write '(("msw" "Microsoft® Word®")
               ("Marquez" "Márquez")
               ("Diaz" "Díaz")
               ("Alvarez" "Álvarez")
               ("Rios" "Ríos")
               ("Lopez" "López")
               ("emacsver" "" (lambda ()
                                (interactive)
                                (emacs-version :here) ))
               ("thedate" "" date)
               ("teln"  "0702541522")
               ("homep" "http://user.it.uu.se/~embe8573")))
      (groups '(("rsb" "rec.sport.boxing")
                ("geh" "gnu.emacs.help")
                ("geg" "gnu.emacs.gnus")))
      (both (append write groups)))
  (define-abbrev-table 'global-abbrev-table both) )

> Gak!  No!  Mother Hitton's Littul Kittons, no!  TEXT
> processor.  I loathe word processores, Word,
> LibreOffice, Abiword, whatever.  They give me the
> screaming heebie-geebies whenever I have to use one
> of them.  Give me Emacs and Groff or TeX.

I agree, that's why I was so surprised to (not) hear
word processor and Groff and LaTeX from the same
person. I never heard "text processor". But, "text
processor" = editor, then?

> But I think we'll have to agree to disagree, here.  I
> am going to investigate Drew's ucs-cmds.el library; it
> looks interesting.  Besides, I think we're getting a
> bit off-topic here.

I always agree to degree in general, only in this case
I'm fairly sure my solution is better, and that this
could be measured and confirmed if a bunch of
scientists did an examination (which I don't suggest
they do, as they should do other thing).

Summary:

For the odd case use: the compose key
For recurrent use in words: abbrev (compare Microsoft®
Word® - and realize that I didn't type that, just m s w
without spaces).

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

RE: .emacs poser

[Drew Adams]

> I see that your using \xABCD notation while I do have Unicode characters
> encoded as \uABCD in my .emacs file.
> 
> I've tried to find information about that in the Emacs manual but saw no
> reference to either \x or \u notations.

(elisp) `General Escape Syntax'

Re: .emacs poser

[Emanuel Berg]

Tilman Ahr <tilman.ahr@mailbox.tu-berlin.de> writes:

> Although I'd *hate* to have to use that to get Ä,Ö,Ü
> and ß. I use those *a lot* when writing in German,
> and the DE-DE layout makes them accessible by a
> single Keystroke.

Well, I certainly hope so :) That is why the local
keymaps are there. The Swedish keymap is great for
writing in Swedish (of course, because of å, ä, and ö)
but as for programming the US layout is 100 times
better.

A couple of friends were telling me this for several
years. I always told them "I know the keys and
shortcuts on the Swedish keyboard. Why would I spend
time learning new?"  But then I tried it *once* and
immediately understood. It is not about "other" keys
and shortcuts, it is about *better* keys and
shortcuts. For example, when doing C, on the US layout,
';' is one key. On the Swedish ditto, it is *two*. You
don't have to write much more than a hello_word.c demo
to realize how much more pleasant it is not having to
do that extra hit at the end of every line. (And this
is just one example: there are also all the brackets.)

> I do have the „dead keys“ option enabled (to make
> typing the usual accents for french, at least,
> feasible without much ado) and remap CAPSLOCK[1] to
> the compose key, nonetheless.

I also remapped caps to

(defun buffer-menu-files-only ()
  (interactive)
  (buffer-menu t) ) ;; files only

It is better, though I could think of a lot to do with
caps that makes more sense than what it is used for by
default. Only time I miss it was for C_CONSTANTS, so I
configured caps-mode.el (which implements buffer-local
caps) to enable on M-caps, and then automatically
disable when you are done typing the const (i.e., for
example when you hit space).

>> Which by the way is another solution that I think is
>> much better than setting this up in Emacs.
>
> Yes. Definitely. But there are (coff, coff) Operating
> Systems that make using a compose key hard, if not
> impossible...

Yeah, like what?

> Depending on the area one works in, and considering
> how the US civil justice system works(fsvo), I could
> see that. Personally, I know how to create © and ® on
> my keyboard, how I'd get the [TM] symbol, I have no
> Idea. I believe I recall it's trivial with a compose
> key, and should be doable without, but I don't need
> that symbol other than for ironic remarks on usenet,
> and there, [TM] does the trick quite satisfactorily.

Yes, again abbrev mode could be setup if you really
want the "real thing".

> So, in everyday use: Yes, it's mainly used as a
> format for well-defined visual representation of …
> data. And in this day and age of the paperless
> office, these are usually printed out.

PDF is for ambitious documents/manuals/books that
require special notation, are likely to be printed, are
unlikely to be changed more than occasionally, and
should be read by humans, and thus should look the same
on all computers. (That's why it is portable (the P), I
guess.) When all (or most) of those properties hold,
you should use it.

For many cases, ASCII is better, because it is
faster. If you did everything in LaTeX, it would take
twice or thrice the time. You would soon spend more
time pedantically making it look hot, forgetting what
you were doing in the first place.

> I had a job interview last week where the interviewer
> had my e-mail with all the relevant stuff for my
> application (CV, Certificates, References and all
> that) as a PDF on his PCs screen right in front of
> him, but still printed it out. On good paper, using a
> colour ink-jet printer that went to photo mode,
> because there was a photo on the CV (most german
> employers really appreciate a CV with a decent
> photo. No idea why).

It the emotional appeal. It doesn't always make
sense. But it is the way humans work, so why fight it?

>> Groff! Wow, you are a man (pun) of many
>> surprises. Is that used outside of the Unix manpages
>> world?
>
> It used to be. And I remember coming across a couple
> of projects where it was used to nicely format simple
> textual data in an automated way even a few years ago

That's interesting, didn't know that.

> Yes. I like abbrev-mode for larger projects. Takes a
> while to get that up to speed, but when it's ready,
> it flies.

Yes!

> Footnotes: [1] I have no idea why anybody would have
> a use for that key's original function on anything
> but a mechanic typewriter. Other than SHOUTING YOUR
> LUNGS OUT…

*And* for #define C_CONSTANT and for non-parameters in
SQL (SELECT year ... ) - though when you think of it,
that convention probably should be dropped. It is just
that your eyes are used it it, so it will be like
teaching an old dog how to sit.

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

Re: .emacs poser

[Emanuel Berg]

"Sebastien Vauban" <sva-news@mygooglest.com> writes:

> I've tried to find information about that in the
> Emacs manual but saw no reference to either \x or \u
> notations.

Isn't x = hex escape (for multibyte non-ASCII > 256),
while u = Unicode, and U = Unicode with more digits?

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

Re: .emacs poser

[Emanuel Berg]

Doug Lewan <dougl@shubertticketing.com> writes:

> I'm surprised that there's only one mention of input
> methods and so much about composing characters, which
> is, as you point out, a little clumsy.

It depends on where the compose key is, and the compose
sequences. If you can do it all without reaching
(i.e. all but the active fingers always at asdf jkl;
positions) it is not really clumsy. The C-x yada yada
yadas are much more clumsy.

But you should only use the compose key for a few
special cases (which will enter the muscle memory, and,
again not be that clumpy if setup close to begin with),
*and* to have the general ability to do whatever
according to some pattern (but now we are talking
*really* rare cases). Of course you should have that
freedom.

Again, take a look at this list:

compose 'o' 'a' to U+00E5 # å
compose 'o' 'A' to U+00C5 # Å
compose '0' 'a' to U+00E5 # å
compose '0' 'A' to U+00C5 # Å
compose '"' 'a' to U+00E4 # ä
compose '"' 'A' to U+00C4 # Ä
compose '/' 'o' to U+00F3 # ó
compose '/' 'O' to U+00D3 # Ó
compose '"' 'o' to U+00F6 # ö
compose '"' 'O' to U+00D6 # Ö
compose '/' 'e' to U+00E9 # é
compose '/' 'E' to U+00C9 # É
compose '/' 'a' to U+00E1 # á
compose '/' 'A' to U+00C1 # Á
compose '/' 'u' to U+00FA # ú
compose '/' 'U' to U+00DA # Ú
compose '/' 'i' to U+00ED # í
compose '/' 'I' to U+00CD # Í
compose '\' 'a' to U+00E0 # à
compose '\' 'A' to U+00C0 # À

Of course, the main thing for me is to be able to use
ä, å, and ö (and their uppercases), but as you see
there is a logical pattern so if I ever have to do á I
can stop, think, and do it.

However, I agree that the compose key is just a tiny
bit clumsy compared to just type, which is why I
prefer to have abbrev deal with whole words in almost
all cases.

> (setq default-input-method "latin-1-prefix") ...

If this works as I think it does, it is potentially
better or at least as good as the compose key solution,
only

- it only works in Emacs
- I'm a bit hesitant if I'd like the prefix to appear
  (as I guess it does?) before it changes the char -
  won't that be disruptive?

But: I'll get back after I try this.

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

Re: .emacs poser

[Emanuel Berg]

Doug Lewan <dougl@shubertticketing.com> writes:

> (setq default-input-method "latin-1-prefix") (let
> ((quail-setup-buf (get-buffer-create "*Quail setup*")))
> (save-excursion (set-buffer quail-setup-buf)
> (toggle-input-method) ;; MAINTENANCE Look for SYMBOLS
> (quail-defrule ":A" ?∀) (quail-defrule ":E" ?∃)
> (quail-defrule ":e" ?∈) ...))

I didn't get it to work. How do you use it?

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

Re: .emacs poser

[Dale Snell]

On Tue, 17 Dec 2013 17:17:16 +0100
Emanuel Berg <embe8573@student.uu.se> wrote:

> I never heard "text processor". But, "text
> processor" = editor, then?

Nope.  A text processor is a stand-alone program; you might call
it a "text assembler".  You write up a source file full of
formatting commands embedded the running text.  You can use any
editor for this, from ed to emacs.  Then you feed that file to the
formatter, and it outputs nicely formatted text, with equations,
figures, and so on, in PS, PDF, HTML, or even straight text.

--Dale

--
"Don't only practice your Art, but force your way into its
Secrets, for it and Knowledge can raise Men to the Divine."
    -- Ludwig van Beethoven

Re: .emacs poser

[Emanuel Berg]

Dale Snell <ddsnell@frontier.com> writes:

>> I never heard "text processor". But, "text processor"
>> = editor, then?
>
> Nope.  A text processor is a stand-alone program ...
> you might call it a "text assembler".  You write up a
> source file full of formatting commands embedded the
> running text.  You can use any editor for this, from
> ed to emacs.  Then you feed that file to the
> formatter, and it outputs nicely formatted text, with
> equations, figures, and so on, in PS, PDF, HTML, or
> even straight text.

As in *markup*? Don't you use Emacs (or another editor)
for this? An editor is a stand-alone program.

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

Re: .emacs poser

[Eli Zaretskii]

> From: "Sebastien Vauban" <sva-news@mygooglest.com>
> Date: Tue, 17 Dec 2013 16:40:38 +0100
> 
> > (global-set-key "\C-cM" (lambda ()
> > 			  "Insert U+2720 = \"✠\"."
> > 			  (interactive)
> > 			  (insert ?\x2720)))
> 
> I see that your using \xABCD notation while I do have Unicode characters
> encoded as \uABCD in my .emacs file.
> 
> I've tried to find information about that in the Emacs manual but saw no
> reference to either \x or \u notations.
> 
> What's the difference?

See the node "General Escape Syntax" in the ELisp manual.

RE: .emacs poser

[Doug Lewan]

Whoops! I omitted an important detail.

You need to be using the latin-1-prefix input method. C-\ will ask you for an input method.

I hope this helps better.

,Douglas
Douglas Lewan
Shubert Ticketing
(201) 489-8600 ext 224

Statistics: The science of figuring out exactly how wrong you probably are.


-----Original Message-----
From: help-gnu-emacs-bounces+dougl=shubertticketing.com@gnu.org [mailto:help-gnu-emacs-bounces+dougl=shubertticketing.com@gnu.org] On Behalf Of Emanuel Berg
Sent: Tuesday, 2013 December 17 12:16
To: help-gnu-emacs@gnu.org
Subject: Re: .emacs poser

Doug Lewan <dougl@shubertticketing.com> writes:

> (setq default-input-method "latin-1-prefix") (let
> ((quail-setup-buf (get-buffer-create "*Quail setup*")))
> (save-excursion (set-buffer quail-setup-buf)
> (toggle-input-method) ;; MAINTENANCE Look for SYMBOLS
> (quail-defrule ":A" ?∀) (quail-defrule ":E" ?∃)
> (quail-defrule ":e" ?∈) ...))

I didn't get it to work. How do you use it?

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

Re: .emacs poser

[Joost Kremers]

Doug Lewan wrote:
> I'm surprised that there's only one mention of input methods and so
> much about composing characters, which is, as you point out, a little
> clumsy.

Yeah, my thoughts exactly.

> It's easy to extend an input method with new rules. I frequently use
> the latin-1-prefix mode for a handful of special characters that I
> use. The lisp to add my rules looks like this:
>
> (setq default-input-method "latin-1-prefix")
> (let ((quail-setup-buf (get-buffer-create "*Quail setup*")))
>   (save-excursion (set-buffer quail-setup-buf)
> 		  (toggle-input-method)
> 		  ;; MAINTENANCE Look for SYMBOLS
> 		  (quail-defrule ":A" ?∀)
> 		  (quail-defrule ":E" ?∃)
> 		  (quail-defrule ":e" ?∈)
>                                      ...))

It's also possible to define completely new input methods. I load the
following code from my init.el:

(quail-define-package
 "arabic-trans" "Ar-trans" "A>" t
 "Input method for Arabic transcription."
 nil t nil nil nil nil nil nil nil nil t)

(quail-define-rules
 (".-" ?ʔ) ;;   U0294 # IPA GLOTTAL STOP
 ("-A" ?Ā) ;;   U0100 # LATIN CAPITAL LETTER A WITH MACRON
 ("-a" ?ā) ;;   U0101 # LATIN SMALL LETTER A WITH MACRON
 ("-D" ?Ḏ) ;;   U1E0E # LATIN CAPITAL LETTER D WITH LINE BELOW
 ("-d" ?ḏ) ;;   U1E0F # LATIN SMALL LETTER D WITH LINE BELOW
 ...)

> I also notice this comment near that code:
> ;; This seems rather a bit of a hack.
> That was my sense long ago; I don't know if that's really a fair sentiment anymore.

Well, one thing that I find strange about defining input methods is that
the functions to define rules do not say to which input method the new
rule should be added. Apparently, quail-defrule adds the rule to the
current input method, while quail-define-rules seems to add them to the
most recently defined one.

Another thing that I find mildly annoying about input methods is that
even though you can set default-input-method, you still need to switch
input methods yourself. Every new buffer starts out without an input
method, so you need to do `C-\` first.

Other than that, I think they're the best way to input non-ASCII
characters.


-- 
Joost Kremers                                   joostkremers@fastmail.fm
Selbst in die Unterwelt dringt durch Spalten Licht
EN:SiS(9)

Re: .emacs poser

[Emanuel Berg]

Doug Lewan <dougl@shubertticketing.com> writes:

> Whoops! I omitted an important detail.
>
> You need to be using the latin-1-prefix input
> method. C-\ will ask you for an input method.

I have C-x RET C-\ for `set-input-method', and C-\ for
`toggle-input-method'.

Yes - then it works exactly as I thought it would. The
thing I don't like (as I mentioned) is that the prefix
char is visible before the translation is done. But I
like it a lot anyway. I don't think I ever use the
Swedish chars in zsh, so I doesn't really matter that
this will be an Emacs only solution. Plus, I still have
the compose key. I think I will start using this as for
now.

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

Re: .emacs poser

[Emanuel Berg]

Joost Kremers <joost.m.kremers@gmail.com> writes:

> Every new buffer starts out without an input method,
> so you need to do `C-\` first.

I should be possible to set the default somehow, or, if
not so, this should be solvable with hooks, the Emacs
poor-man's solution for everything when everything else
fails.

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

Re: .emacs poser

[Emanuel Berg]

Emanuel Berg <embe8573@student.uu.se> writes:

> I should be possible to set the default somehow, or,
> if not so, this should be solvable with hooks, the
> Emacs poor-man's solution for everything when
> everything else fails.

I got it working like this. It will be interesting to
see if it works out.

(quail-define-package "swedish-quail" "Swedish" "Swedish quail")

(quail-define-rules
 ("0A" ?Å) ("0a" ?å)
 (":A" ?Ä) (":a" ?ä)
 (":O" ?Ö) (":o" ?ö) )
(set-input-method 'swedish-quail) ;; add this to text-mode-hook etc.

In the docstring to the `quail-define-package', there
is an optional argument described like this:

;; GUIDANCE specifies how a guidance string is shown in
;; echo area.  If it is t, list of all possible
;; translations for the current key is shown with the
;; currently selected translation being highlighted.
;; If it is an alist, the element has the form (CHAR
;; . STRING).  Each character in the current key is
;; searched in the list and the corresponding string is
;; shown.  If it is nil, the current key is shown.

However, is it not possible to entirely disable the
echo area fanfare? I think it is just disruptive.

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

RE: .emacs poser

[Doug Lewan]

> Yes - then it works exactly as I thought it would. The
> thing I don't like (as I mentioned) is that the prefix
> char is visible before the translation is done.

One of the early requirements for supporting multiple character sets was that it be implemented in a single, unifying framework. What you don't like is a consequence of that.

European character sets tend to be simple: Roman alphabet + modifications. Many languages are much more involved, and the input method correspondingly so. TeX as an input method can give you a hint about why you might want to see the input-so-far: You can proofread it as you go.

,Douglas
Douglas Lewan
Shubert Ticketing
(201) 489-8600 ext 224

Statistics: The science of figuring out exactly how wrong you probably are.


> -----Original Message-----
> From: help-gnu-emacs-bounces+dougl=shubertticketing.com@gnu.org
> [mailto:help-gnu-emacs-bounces+dougl=shubertticketing.com@gnu.org] On
> Behalf Of Emanuel Berg
> Sent: Tuesday, 2013 December 17 17:42
> To: help-gnu-emacs@gnu.org
> Subject: Re: .emacs poser
> 
> Doug Lewan <dougl@shubertticketing.com> writes:
> 
> > Whoops! I omitted an important detail.
> >
> > You need to be using the latin-1-prefix input
> > method. C-\ will ask you for an input method.
> 
> I have C-x RET C-\ for `set-input-method', and C-\ for
> `toggle-input-method'.
> 
> Yes - then it works exactly as I thought it would. The
> thing I don't like (as I mentioned) is that the prefix
> char is visible before the translation is done. But I
> like it a lot anyway. I don't think I ever use the
> Swedish chars in zsh, so I doesn't really matter that
> this will be an Emacs only solution. Plus, I still have
> the compose key. I think I will start using this as for
> now.
> 
> --
> Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
> underground experts united:  http://user.it.uu.se/~embe8573

Re: .emacs poser

[Yuri Khan]

On Tue, Dec 17, 2013 at 10:17 PM, Drew Adams <drew.adams@oracle.com> wrote:

>> IMO you’re doing it wrong, by solving the general problem of entering
>> special character for a specific case of Emacs.
>
> There is no wrong or right way here.  Does your keyboard have a key
> that inserts the character `2' (yes)?  Is dedicating that key
> sequence to inserting that one character wasteful or "doing it wrong"?

No, because this works in the VTs, in all programs running in VTs, on
all X displays, and in all programs running on X displays.

> Binding a command that inserts a commonly used character to a key
> can be quite handy, regardless of whether there is also a general
> way to insert any character.

Exactly. But I argue that if one wants a special character in Emacs,
one will also want the same character in other programs, entered by
the same means. Therefore, the place to set up that binding is the XKB
map and the corresponding console map.

Re: .emacs poser

[Tilman Ahr]

Emanuel Berg <embe8573@student.uu.se> writes:

> Tilman Ahr <tilman.ahr@mailbox.tu-berlin.de> writes:

[...]

> but as for programming the US layout is 100 times
> better.
>
> A couple of friends were telling me this for several
> years. I always told them "I know the keys and
> shortcuts on the Swedish keyboard. Why would I spend
> time learning new?"  But then I tried it *once* and
> immediately understood. It is not about "other" keys
> and shortcuts, it is about *better* keys and
> shortcuts. For example, when doing C, on the US layout,
> ';' is one key. On the Swedish ditto, it is *two*. You
> don't have to write much more than a hello_word.c demo
> to realize how much more pleasant it is not having to
> do that extra hit at the end of every line. (And this
> is just one example: there are also all the brackets.)

I've never really viewed shifting as a seperate keystroke, but then I've
never done whole lots of programming…

>> I do have the „dead keys“ option enabled (to make
>> typing the usual accents for french, at least,
>> feasible without much ado) and remap CAPSLOCK[1] to
>> the compose key, nonetheless.
>
> I also remapped caps to
>
> (defun buffer-menu-files-only ()
>   (interactive)
>   (buffer-menu t) ) ;; files only

Ok. I like having it as a compose key, because I need that somewhat
frequently. It's done on a OS (GUI, really) level, so it works
consistently in all my applications. I don't really use one so
overwhelmingly much mort than the rest that it would make sense to have
application-specific custon key-mapping. Might be different if I
actually used my PC for work.

>> Yes. Definitely. But there are (coff, coff) Operating
>> Systems that make using a compose key hard, if not
>> impossible...
>
> Yeah, like what?

I never found a sane way to do it in Windows, for example. Might have
gotten better in the last couple of years, though. My experience with
Windows has been a while ago.

[...]

>>> Groff! Wow, you are a man (pun) of many
>>> surprises. Is that used outside of the Unix manpages
>>> world?
>>
>> It used to be. And I remember coming across a couple
>> of projects where it was used to nicely format simple
>> textual data in an automated way even a few years ago
>
> That's interesting, didn't know that.

IIRC the rationale was that it did all the formatting  needed for the
task at hand, and was significantly simpler and less resource-hungry
than LaTeX & Co.

[...]

>> Footnotes: [1] I have no idea why anybody would have
>> a use for that key's original function on anything
>> but a mechanic typewriter. Other than SHOUTING YOUR
>> LUNGS OUT…
>
> *And* for #define C_CONSTANT and for non-parameters in
> SQL (SELECT year ... ) - 

Alway used to just use the regular shift key for stuff like that, but
then, as I stated earlier, I've never spend significant time
programming. If you'd need to do a lot of that, CapsLock would surely be
nice to have.

> though when you think of it, that convention probably should be
> dropped.

There are quite a few conventions like that, when you really think about
it. 

> It is just that your eyes are used it it, so it will be like teaching
> an old dog how to sit.

Something like that, yes.

Re: .emacs poser

[Yuri Khan]

On Tue, Dec 17, 2013 at 11:42 PM, Emanuel Berg <embe8573@student.uu.se> wrote:

>> Footnotes: [1] I have no idea why anybody would have
>> a use for that key's original function on anything
>> but a mechanic typewriter. Other than SHOUTING YOUR
>> LUNGS OUT…
>
> *And* for #define C_CONSTANT and for non-parameters in
> SQL (SELECT year ... ) - though when you think of it,
> that convention probably should be dropped. It is just
> that your eyes are used it it, so it will be like
> teaching an old dog how to sit.

Writing keywords in all caps used to be popular in case-insensitive
languages until syntax highlighting editors became the norm. Nowadays,
it’s not quite as frequent.

If/when there comes a day when most editors (IDEs?) can reliably
detect preprocessor macros and highlight them with a dangerous-looking
shade of red, the all-caps convention for #defines can also be
dropped.

Re: .emacs poser

[Jambunathan K]

Joost Kremers <joost.m.kremers@gmail.com> writes:

> Every new buffer starts out without an input method, so you need to do
> `C-\` first.

If you do a

   M-x set-language-environment

either interactively or in your .emacs, then you can switch between
input methods with just C-\.  You don't have to "select" an input
method.  Some input method apropriate for the language gets chosen
apropriately.  I am not sure "how" the default input method is chosen
but the fact is it gets chosen.

Re: .emacs poser

[Emanuel Berg]

Tilman Ahr <tilman.ahr@mailbox.tu-berlin.de> writes:

> I've never really viewed shifting as a seperate
> keystroke, but then I've never done whole lots of
> programming...

Like I said, I couldn't understand it at all until I
tried it. Then it instantly made sense.

For C programming, just think where the ';' is. It is
at the little finger of the right hand, as is all the
brackets. Those are used at the end of lines. So from
there, you just move two keys to the right and hit RET
to get a new line.

But - "Unfortunately, no one can be 'told' what the
Matrix is. You have to see it for yourself."

> Ok. I like having it as a compose key, because I need
> that somewhat frequently. It's done on a OS (GUI,
> really) level, so it works consistently in all my
> applications.

Yes, and this is the appeal of the compose key.

> My experience with Windows has been a while ago.

You old bragger!

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

Re: .emacs poser

[Emanuel Berg]

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

> Writing keywords in all caps used to be popular in
> case-insensitive languages until syntax highlighting
> editors became the norm. Nowadays, it’s not quite as
> frequent.

Is the preprocessor in C case-*in*sensitive? Because C
isn't.

> If/when there comes a day when most editors (IDEs?)
> can reliably detect preprocessor macros and highlight
> them with a dangerous-looking shade of red, the
> all-caps convention for #defines can also be dropped.

This sounds like something that would be possible to do
with Emacs today. As for the definitions, that would be
a regular expression (very easy). But how to make it
consistent in the rest of the buffer (i.e., where the
constant is used), I don't know.

In c-mode:

#ifndef DEF_H
#define FINGERS 10

The faces are:

#ifndef, #define: font-lock-preprocessor-face
FINGERS: font-lock-variable-name-face

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

Re: .emacs poser

[Emanuel Berg]

Jambunathan K <kjambunathan@gmail.com> writes:

>> Every new buffer starts out without an input method,
>> so you need to do `C-\` first.
>
> If you do a
>
>    M-x set-language-environment
>
> either interactively or in your .emacs, then you can
> switch between input methods with just C-\.

How is that any different from what he said? He said
you have to do a C-\, and you say you can do it with:
C-\.

> You don't have to "select" an input method.

How is that not selecting?

If I do (to sync with the definition I provided
earlier, I don't know if there is an official Swedish
setup):

(set-language-environment "Swedish")

I still have to do C-\. However, this is possible to
not have to do with hooks. (And I highly recommend not
doing it every time.)

(add-hook 'text-mode-hook
          (lambda ()
            (set-input-method 'swedish-quail)
            ;; ...
            ))

But, there seems to be more than meets the eye. In this
buffer, in message-mode, it works, and I didn't setup
that hook...

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

Re: .emacs poser

[Yuri Khan]

On Thu, Dec 19, 2013 at 4:23 AM, Emanuel Berg <embe8573@student.uu.se> wrote:

>> Writing keywords in all caps used to be popular in
>> case-insensitive languages until syntax highlighting
>> editors became the norm. Nowadays, it’s not quite as
>> frequent.
>
> Is the preprocessor in C case-*in*sensitive? Because C
> isn't.

The preprocessor is case-sensitive, as is C. But, contrary to language
keywords, we the developers get to pick our macro names.

Because macro names are not namespaced and because they are known to
cause surprising effects when used carelessly, we pick them so that
they be highly visible and less likely to clash with other
identifiers. To the latter end, ALL CAPS provides a mini-namespace, if
we agree never to name non-macros in all caps.

>> If/when there comes a day when most editors (IDEs?)
>> can reliably detect preprocessor macros and highlight
>> them with a dangerous-looking shade of red, the
>> all-caps convention for #defines can also be dropped.
>
> This sounds like something that would be possible to do
> with Emacs today. As for the definitions, that would be
> a regular expression (very easy). But how to make it
> consistent in the rest of the buffer (i.e., where the
> constant is used), I don't know.

The highlight at definition time is not important; there we have the #
character to alert us that something shady is going on.

What’s important is the highlight at each use. Syntactically, macros
do not differ from ordinary identifiers, and their set is not constant
as in the case of language keywords. So the IDE must do some parsing
and remembering.

Also, macro definitions may be placed under conditional compilation.
So, depending on which branch is taken, different macros may be in
effect after definition. So the IDE must also understand conditional
compilation directives and be able to evaluate conditions.

Additionally, sometimes preprocessor macros are defined in the
compiler invocation command (-D switch for most compilers).


But sorry for off-topic; we are digressing here from the problem of
the rogue macron into the land of semantic parsers.

Re: .emacs poser

[Emanuel Berg]

Jambunathan K <kjambunathan@gmail.com> writes:

> My language is "Tamil" (language: "ta").  It has two
> input methods.  With the above snippet installed, I
> can just C-\ blindly.  Otherwise, I need to input a
> default choice for input method.

OK, so you have several input methods for one language?
Is it old and modern Tamil?

Then toggling perhaps makes sense, I don't know. It
can't visualize that situation, but explain if you
care to, it is interesting.

I have another situation, namely I prefer -

0 + a = å
: + a = ä
: + o = ö

- etc. to be active all the time. Though I type mostly
in English I type in Swedish often enough not to want
to enable/disable it each time.

I haven't used it long enough to be able to say if it
interferes with "normal" typing somehow. The :keywords
in Elisp (if they start with 'a' or 'o'), perhaps. But
for elisp-mode, I can add a hook to disable it - I
never do Elisp in Swedish.

To have it "always on" is part of the point to it all -
this (input) method, and the compose key solution - if
I *did* want to enable it back and forth, I might as
well change the entire layout of the keyboard.

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

Re: .emacs poser

[Emanuel Berg]

Emanuel Berg <embe8573@student.uu.se> writes:

> I haven't used it long enough to be able to say if it
> interferes with "normal" typing somehow. The
> :keywords in Elisp (if they start with 'a' or 'o'),
> perhaps. But for elisp-mode, I can add a hook to
> disable it - I never do Elisp in Swedish.

Come to think of it, it is much better to do it the
other way around. To enable it for the literary modes,
because there are so many programming modes.

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

Re: .emacs poser

[Joost Kremers]

Jambunathan K wrote:
> Joost Kremers <joost.m.kremers@gmail.com> writes:
>
>> Every new buffer starts out without an input method, so you need to do
>> `C-\` first.
>
> If you do a
>
>    M-x set-language-environment
>
> either interactively or in your .emacs, then you can switch between
> input methods with just C-\.

I set the variable default-input-method in my init.el, which allows me
to do the same thing. The point is that if you create a new buffer, it
doesn't have any input method. I have to hit C-\ first to select the
default one.



-- 
Joost Kremers                                   joostkremers@fastmail.fm
Selbst in die Unterwelt dringt durch Spalten Licht
EN:SiS(9)

Re: .emacs poser

[Emanuel Berg]

Joost Kremers <joost.m.kremers@gmail.com> writes:

> I set the variable default-input-method in my
> init.el, which allows me to do the same thing. The
> point is that if you create a new buffer, it doesn't
> have any input method.

That's the point, but...

> I have to hit C-\ first to select the default one.

... you don't *have to* do anything, as you can put
this in hooks.

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

Re: .emacs poser

[Emanuel Berg]

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

> What's important is the highlight at each
> use. Syntactically, macros do not differ from
> ordinary identifiers, and their set is not constant
> as in the case of language keywords. So the IDE must
> do some parsing and remembering.

Check out the thread "random color of variables in
programming languages" and this dump [1]. Shouldn't
what you describe be possible to do with that same
method?

[1] http://user.it.uu.se/~embe8573/colvar.png

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

Re: .emacs poser

[Jambunathan K]

Emanuel Berg <embe8573@student.uu.se> writes:

> (set-language-environment "Swedish")

What happens if you add the above snippet to your .emacs?

My language is "Tamil" (language: "ta").  It has two input methods.
With the above snippet installed, I can just C-\ blindly.  Otherwise, I
need to input a default choice for input method.

Re: .emacs poser

[Jambunathan K]

Emanuel Berg <embe8573@student.uu.se> writes:

> 0 + a = å
> : + a = ä
> : + o = ö

If set your input method to

    latin-1-prefix

you can use /a, :a, :o to input the above characters.

Why is it that you want roll your own?

Btw, I see a swedish-postfix input method.  What you want is
swedish-prefix input method.  Seems like a good candidate for M-x
report-emacs-bug.  If you propose a patch, then you can get the prefixes
that you yourself use.

Re: .emacs poser

[Jambunathan K]

Jambunathan K <kjambunathan@gmail.com> writes:

>> 0 + a = å
>> : + a = ä
>> : + o = ö
>
> If set your input method to
>
>     latin-1-prefix
>
> you can use /a, :a, :o to input the above characters.
                  ^^  ^^
                  "a  "o

Btw, if you want the input sequence verbatim you can use the C-q trick.
So C-q / a will give you /a verbatim in the buffer.

Re: .emacs poser

[Emanuel Berg]

Jambunathan K <kjambunathan@gmail.com> writes:

>> 0 + a = å : + a = ä : + o = ö
>
> If set your input method to
>
>     latin-1-prefix
>
> you can use /a, :a, :o to input the above characters.
>
> Why is it that you want roll your own?

In the post, I describe how it differs:

1. [compose] har = h [compose] ar = här

2. The prefix/compose char is not inserted: you only
   see the letters that are part of the word. (Compare:
   you could setup an abbrev to make "nasa" into
   NASA. But I don't want NASA to look like "nasa",
   ever.)
   
3. No visual noise in the echo area.

Also, though a minor matter, I can set it up on a
mode-by-mode basis without having to use the poor man's
hooks.

> Btw, I see a swedish-postfix input method.  What you
> want is swedish-prefix input method.

That was my first attempt:

(quail-define-package "swedish-quail" "Swedish" "Swedish quail")
(quail-define-rules
 ("0A" ?Å) ("0a" ?å)
 (":A" ?Ä) (":a" ?ä)
 (":O" ?Ö) (":o" ?ö) )
(set-input-method 'swedish-quail)

However (again as I said even back then) I don't like
the insertion of the prefix before it transforms, and I
don't like the visual noise in the echo area.

> Seems like a good candidate for M-x report-emacs-bug.

Or a good candidate for writing that *gorgeous*
Elisp. Red hair, green eyes, soft skin...

> If you propose a patch, then you can get the prefixes
> that you yourself use.

I think this as a patch would have had a hard time as
most people probably want the semi-colon for natural
languages.

-- 
Emanuel Berg, programmer-for-rent. CV, projects, etc at uXu
underground experts united:  http://user.it.uu.se/~embe8573

Re: .emacs poser

[B. T. Raven]

Emacsers:

Thanks for all the answers. After returning from a 10 day trip I thought
I might get a couple of responses but there were enough that two or
three of them solved the problem. There was an extra glyph (combining
macron or modifier letter macron) hidden in the question mark (character
declaration). I use this method just to experiment with utf-8 and
usually insert most exotic characters with input methods (usually
latin-postfix). I know about ucs-insert but don't want to have to look
up the hexcodes. Apparently once I used C-c M for combining macron and
then for Maltese cross. Usually I do C-x C-e just on the changed line to
test but I must have had this bug for a long time and then just
commented it out many months ago instead of digging into the problem.

Thanks again.

Ed

> Jambunathan K <kjambunathan@gmail.com> writes:
> 
>>> 0 + a = å : + a = ä : + o = ö
>>
>> If set your input method to
>>
>>     latin-1-prefix
>>
>> you can use /a, :a, :o to input the above characters.
>>
>> Why is it that you want roll your own?
> 
> In the post, I describe how it differs:
> 
> 1. [compose] har = h [compose] ar = här
> 
> 2. The prefix/compose char is not inserted: you only
>    see the letters that are part of the word. (Compare:
>    you could setup an abbrev to make "nasa" into
>    NASA. But I don't want NASA to look like "nasa",
>    ever.)
>    
> 3. No visual noise in the echo area.
> 
> Also, though a minor matter, I can set it up on a
> mode-by-mode basis without having to use the poor man's
> hooks.
> 
>> Btw, I see a swedish-postfix input method.  What you
>> want is swedish-prefix input method.
> 
> That was my first attempt:
> 
> (quail-define-package "swedish-quail" "Swedish" "Swedish quail")
> (quail-define-rules
>  ("0A" ?Å) ("0a" ?å)
>  (":A" ?Ä) (":a" ?ä)
>  (":O" ?Ö) (":o" ?ö) )
> (set-input-method 'swedish-quail)
> 
> However (again as I said even back then) I don't like
> the insertion of the prefix before it transforms, and I
> don't like the visual noise in the echo area.
> 
>> Seems like a good candidate for M-x report-emacs-bug.
> 
> Or a good candidate for writing that *gorgeous*
> Elisp. Red hair, green eyes, soft skin...
> 
>> If you propose a patch, then you can get the prefixes
>> that you yourself use.
> 
> I think this as a patch would have had a hard time as
> most people probably want the semi-colon for natural
> languages.
> 

Re: .emacs poser

[Yuri Khan]

On Fri, Dec 27, 2013 at 1:00 PM, B. T. Raven <btraven@nihilo.net> wrote:

> I know about ucs-insert but don't want to have to look
> up the hexcodes.

You do not have to know the hex code — the canonical character name
will do just fine, and there is tab completion.

RE: .emacs poser

[Drew Adams]

> > I know about ucs-insert but don't want to have to look
> > up the hexcodes.
> 
> You do not have to know the hex code — the canonical character name
> will do just fine, and there is tab completion.

And if you have completion with regexp or substring patterns
then it becomes even simpler.

Re: .emacs poser

[B. T. Raven]

Wow! Who'd'a thunk? I entered in lower case 'combining', tab upcased it,
and tab again gave a whole page of canonical names starting with COMBINING
!!

>>> I know about ucs-insert but don't want to have to look
>>> up the hexcodes.
>>
>> You do not have to know the hex code — the canonical character name
>> will do just fine, and there is tab completion.
> 
> And if you have completion with regexp or substring patterns
> then it becomes even simpler.

Any regular expression in the mini-buffer returns "Not a unicode
character..."


Ed
> 

RE: .emacs poser

[Drew Adams]

> >>> I know about ucs-insert but don't want to have to look
> >>> up the hexcodes.
> >>
> >> You do not have to know the hex code — the canonical character name
> >> will do just fine, and there is tab completion.
>
> Wow!  Who'd'a thunk? I entered in lower case 'combining', tab
> upcased it, and tab again gave a whole page of canonical names
> starting with COMBINING !!
> 
> > And if you have completion with regexp or substring patterns
> > then it becomes even simpler.
> 
> Any regular expression in the mini-buffer returns "Not a unicode
> character..."

I said, "IF you have completion with regexp or substring patterns".

If you use Icicles, for instance, you can match using such patterns.
And you can use progressive completion: match any number of such
patterns, in any order (e.g., multiple substrings).

Actually, you can even match against the character itself.  Why
might you want to do that?  To see the corresponding Unicode
character name(s), including any old names.  For example, for
the character ` you get these two completion candidates:

 GRAVE ACCENT	`
 SPACING GRAVE	`

Completion candidates for this command are multi-completions: the
char name followed by the char itself, separated by a TAB char.

The main purpose for this is to show you the characters along
with their names (WYSIWYG).  The chars are shown in
*Completions* using the same font as the frame from which you
invoked `C-x 8 RET'.  That you can also match against the chars
themselves is just a nice-to-have.

Another part of this feature: When you cycle among the
candidates that match your input, the mode line of *Completions*
shows the char together with its code point (in hex, octal, and
decimal).

The most important parts of this feature, however, are (a)
seeing the characters (WYSIWYG) and (b) being able to use
progressive completion, so you can use multiple simple patterns
instead of a single complex pattern.