How to redefine Parantheses?

How to redefine Parantheses?

[Paul Batt]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 288 bytes --]

I need German quotation marks « and » (ISO 8859 -1 Latin-1 171 and 187). 
Alas, Emacs treats them as parantheses, which is sometimes disturbing. Is 
there a way I could redefine what Emacs considers parantheses in order to 
have those quotation marks treated as normal text?
Thanks, Paul

Re: How to redefine Parantheses?

[Eli Zaretskii]

> From: "Paul Batt" <paul.batt@gmx.net>
> Date: Sun, 8 Apr 2007 09:19:12 +0200
> 
> I need German quotation marks « and » (ISO 8859 -1 Latin-1 171 and 187). 
> Alas, Emacs treats them as parantheses, which is sometimes disturbing. Is 
> there a way I could redefine what Emacs considers parantheses in order to 
> have those quotation marks treated as normal text?

You need to change the syntax class of those characters.  See the
documentation of char-syntax and modify-syntax-entry.

Re: How to redefine Parantheses?

[Ralf Angeli]

* Eli Zaretskii (2007-04-08) writes:

>> From: "Paul Batt" <paul.batt@gmx.net>
>> Date: Sun, 8 Apr 2007 09:19:12 +0200
>> 
>> I need German quotation marks « and » (ISO 8859 -1 Latin-1 171 and 187). 
>> Alas, Emacs treats them as parantheses, which is sometimes disturbing. Is 
>> there a way I could redefine what Emacs considers parantheses in order to 
>> have those quotation marks treated as normal text?
>
> You need to change the syntax class of those characters.

No, the default syntax in Emacs has to be changed.

-- 
Ralf

Re: How to redefine Parantheses?

[Eli Zaretskii]

> From: Ralf Angeli <dev.null@caeruleus.net>
> Date: Sun, 08 Apr 2007 11:35:08 +0200
> 
> >> I need German quotation marks « and » (ISO 8859 -1 Latin-1 171 and 187). 
> >> Alas, Emacs treats them as parantheses, which is sometimes disturbing. Is 
> >> there a way I could redefine what Emacs considers parantheses in order to 
> >> have those quotation marks treated as normal text?
> >
> > You need to change the syntax class of those characters.
> 
> No, the default syntax in Emacs has to be changed.

????

Re: How to redefine Parantheses?

[Ralf Angeli]

* Eli Zaretskii (2007-04-08) writes:

>> From: Ralf Angeli <dev.null@caeruleus.net>
>> Date: Sun, 08 Apr 2007 11:35:08 +0200
>> 
>> >> I need German quotation marks « and » (ISO 8859 -1 Latin-1 171 and 187). 
>> >> Alas, Emacs treats them as parantheses, which is sometimes disturbing. Is 
>> >> there a way I could redefine what Emacs considers parantheses in order to 
>> >> have those quotation marks treated as normal text?
>> >
>> > You need to change the syntax class of those characters.
>> 
>> No, the default syntax in Emacs has to be changed.
>
> ????

Well, the default syntax is obviously wrong since you cannot rely on
guillemets being used parenthetical in a certain order.

-- 
Ralf

Re: How to redefine Parantheses?

[Eli Zaretskii]

> From: Ralf Angeli <dev.null@caeruleus.net>
> Date: Sun, 08 Apr 2007 23:15:23 +0200
> 
> >> > You need to change the syntax class of those characters.
> >> 
> >> No, the default syntax in Emacs has to be changed.
> >
> > ????
> 
> Well, the default syntax is obviously wrong since you cannot rely on
> guillemets being used parenthetical in a certain order.

In many (most?) languages, the current syntax does make sense.

Anyway, this kind of discussion doesn't belong here.  I just tried to
help the OP to solve the immediate problem.

Re: How to redefine Parantheses?

[David Hansen]

On Mon, 09 Apr 2007 08:56:31 +0300 Eli Zaretskii wrote:

>> From: Ralf Angeli <dev.null@caeruleus.net>
>> Date: Sun, 08 Apr 2007 23:15:23 +0200
>> 
>> >> > You need to change the syntax class of those characters.
>> >> 
>> >> No, the default syntax in Emacs has to be changed.
>> >
>> > ????
>> 
>> Well, the default syntax is obviously wrong since you cannot rely on
>> guillemets being used parenthetical in a certain order.
>
> In many (most?) languages, the current syntax does make sense.
>
> Anyway, this kind of discussion doesn't belong here.  I just tried to
> help the OP to solve the immediate problem.

AFAIK the discussion already was on the emacs-devel list.

David

Re: How to redefine Parantheses?

[Ralf Angeli]

* David Hansen (2007-04-09) writes:

> On Mon, 09 Apr 2007 08:56:31 +0300 Eli Zaretskii wrote:
>
>> Anyway, this kind of discussion doesn't belong here.  I just tried to
>> help the OP to solve the immediate problem.
>
> AFAIK the discussion already was on the emacs-devel list.

One thread about it I remember was on emacs-pretest-bug:
<URL:http://thread.gmane.org/rzqd5u9irrz.fsf%40dpcsport.dl.ac.uk>

-- 
Ralf

Re: How to redefine Parantheses?

[Ralf Angeli]

* Eli Zaretskii (2007-04-09) writes:

>> From: Ralf Angeli <dev.null@caeruleus.net>
>> Date: Sun, 08 Apr 2007 23:15:23 +0200
>> 
>> >> No, the default syntax in Emacs has to be changed.
>> >
>> > ????
>> 
>> Well, the default syntax is obviously wrong since you cannot rely on
>> guillemets being used parenthetical in a certain order.
>
> In many (most?) languages, the current syntax does make sense.

I just looked at the csquotes LaTeX package which deals with quotation
marks in different languages.  The author made a survey about the use
of quotation marks in different languages when he implemented the
package.  Here are results of using it with different language
options:

»danish«
« french »
»german«
«german (swiss variant)»
«italian»
»swedish»

So there is no clear indication of which variant is more common.  In
Swedish guillemets aren't even used parenthetical.  I really don't
understand why the default syntax in Emacs is paren syntax.
Guillemets are quotation marks which are not consistently used in a
parenthetical way, so using paren syntax is clearly a bug.  And I
don't understand why one does not simply give them punctuation syntax
in order to get rid of the bug and stop this issue from popping up
again and again.

> Anyway, this kind of discussion doesn't belong here.  I just tried to
> help the OP to solve the immediate problem.

Well, I tried so too.  But I don't think is a customization issue but
a bug in Emacs.

-- 
Ralf

Re: How to redefine Parantheses?

[Eli Zaretskii]

> From: Ralf Angeli <dev.null@caeruleus.net>
> Date: Mon, 09 Apr 2007 10:58:35 +0200
> 
> I really don't understand why the default syntax in Emacs is paren
> syntax.

Because Emacs currently cannot distinguish between different usage of
the same character in different languages, and even your research
indicates that the current behavior is correct in 50% of cases.

> But I don't think is a customization issue but a bug in Emacs.

Not a bug, a missing feature.

Re: How to redefine Parantheses?

[Ralf Angeli]

* Eli Zaretskii (2007-04-09) writes:

>> From: Ralf Angeli <dev.null@caeruleus.net>
>> Date: Mon, 09 Apr 2007 10:58:35 +0200
>> 
>> I really don't understand why the default syntax in Emacs is paren
>> syntax.
>
> Because Emacs currently cannot distinguish between different usage of
> the same character in different languages, and even your research
> indicates that the current behavior is correct in 50% of cases.
>
>> But I don't think is a customization issue but a bug in Emacs.
>
> Not a bug, a missing feature.

Even with that feature missing one could get 100% correct behavior by
giving guillemets punctuation syntax.  You may give the problem a
different name, but changing the default syntax is ceteris paribus
still the better option.

-- 
Ralf

Re: How to redefine Parantheses?

[Eli Zaretskii]

> From: Ralf Angeli <dev.null@caeruleus.net>
> Date: Mon, 09 Apr 2007 19:37:36 +0200
> 
> Even with that feature missing one could get 100% correct behavior by
> giving guillemets punctuation syntax.  You may give the problem a
> different name, but changing the default syntax is ceteris paribus
> still the better option.

Ah, but other things are not really the same, since I don't regard the
behavior you want as 100% correct.

Re: How to redefine Parantheses?

[Ralf Angeli]

* Eli Zaretskii (2007-04-10) writes:

>> From: Ralf Angeli <dev.null@caeruleus.net>
>> Date: Mon, 09 Apr 2007 19:37:36 +0200
>> 
>> Even with that feature missing one could get 100% correct behavior by
>> giving guillemets punctuation syntax.  You may give the problem a
>> different name, but changing the default syntax is ceteris paribus
>> still the better option.
>
> Ah, but other things are not really the same, since I don't regard the
> behavior you want as 100% correct.

Please be more precise.  Are you referring to movement commands?

-- 
Ralf

Re: How to redefine Parantheses?

[Eli Zaretskii]

> From: Ralf Angeli <dev.null@caeruleus.net>
> Date: Tue, 10 Apr 2007 19:35:53 +0200
> 
> > Ah, but other things are not really the same, since I don't regard the
> > behavior you want as 100% correct.
> 
> Please be more precise.  Are you referring to movement commands?

That, and the paren-matching features (e.g., show-paren-mode).

Re: How to redefine Parantheses?

[Ralf Angeli]

* Eli Zaretskii (2007-04-10) writes:

>> From: Ralf Angeli <dev.null@caeruleus.net>
>> Date: Tue, 10 Apr 2007 19:35:53 +0200
>> 
>> > Ah, but other things are not really the same, since I don't regard the
>> > behavior you want as 100% correct.
>> 
>> Please be more precise.  Are you referring to movement commands?
>
> That, and the paren-matching features (e.g., show-paren-mode).

We are running round in circles. (c:  Those features fail for 50% of
the users and leave them irritated.  If the default syntax is changed
to punctuation, there will be no irritated users and those who want
paren-related movement and matching can enable the feature by setting
the syntax to _their_ preference.  So they could manually emulate part
of the feature you describe as missing.

FWIW, in AUCTeX we've set the syntax of guillemets to punctuation
since 2004 and since then we haven't got a single complaint about
missing or mislead paren matching.

-- 
Ralf

Re: How to redefine Parantheses?

[Eli Zaretskii]

> From: Ralf Angeli <dev.null@caeruleus.net>
> Date: Tue, 10 Apr 2007 21:50:39 +0200
> 
> * Eli Zaretskii (2007-04-10) writes:
> 
> >> From: Ralf Angeli <dev.null@caeruleus.net>
> >> Date: Tue, 10 Apr 2007 19:35:53 +0200
> >> 
> >> > Ah, but other things are not really the same, since I don't regard the
> >> > behavior you want as 100% correct.
> >> 
> >> Please be more precise.  Are you referring to movement commands?
> >
> > That, and the paren-matching features (e.g., show-paren-mode).
> 
> We are running round in circles.

Maybe you are; I ain't.  You asked me to explain what behavior I like
about the current syntax, and I answered.  It is okay to disagree;
there's certainly no need to reiterate why you dislike the current
behavior, because you already stated that.

Re: How to redefine Parantheses?

[Ralf Angeli]

* Eli Zaretskii (2007-04-11) writes:

>> From: Ralf Angeli <dev.null@caeruleus.net>
>> Date: Tue, 10 Apr 2007 21:50:39 +0200
>> 
>> We are running round in circles.
>
> Maybe you are; I ain't.  You asked me to explain what behavior I like
> about the current syntax, and I answered.  It is okay to disagree;
> there's certainly no need to reiterate why you dislike the current
> behavior, because you already stated that.

I tried to clarify why changing the default syntax is better than
leaving it.  In addition I provided information on what to expect from
changing the syntax by example of AUCTeX.  Please don't just ignore
that.

-- 
Ralf

Re: How to redefine Parantheses?

[Eli Zaretskii]

> From: Ralf Angeli <dev.null@caeruleus.net>
> Date: Wed, 11 Apr 2007 18:53:49 +0200
> 
> I tried to clarify why changing the default syntax is better than
> leaving it.  In addition I provided information on what to expect from
> changing the syntax by example of AUCTeX.  Please don't just ignore
> that.

I heard your opinions loud and clear, and didn't ignore them.  I just
don't agree, and see no way of convincing you; that's why I didn't
respond to that specific part.  Please respect my dissenting views,
and please don't interpret disagreement as rudeness.

Re: How to redefine Parantheses?

[Paul Batt]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1013 bytes --]

"Eli Zaretskii" <eliz@gnu.org> schrieb im Newsbeitrag 
news:mailman.1835.1176020435.7795.help-gnu-emacs@gnu.org...

> You need to change the syntax class of those characters.  See the
> documentation of char-syntax and modify-syntax-entry.

Thank you. I am almost totally lisp illiterate, but after some reading in 
the manual I put

(modify-syntax-entry ?« "." text-mode-syntax-table)
(modify-syntax-entry ?» "." text-mode-syntax-table)

in my .emacs, which in part appears to do what I wanted. Characters « and » 
are no longer treated as parantheses.

However, there is still another problem with these. I also wanted to set up 
a keyboard macro for them. So I put

(fset 'anfein "«")
(global-set-key "\C-z" 'anfein)

in my .emacs. This will work for all characters or strings, but does not for 
« and ». I get an error message, "After 0 kbd macro iterations: Keyboard 
macro terminated by a command ringing the bell" instead.

So something still is weird regarding « and ». What could it be?

Thank you, Paul

Re: How to redefine Parantheses?

[Eli Zaretskii]

> From: "Paul Batt" <paul.batt@gmx.net>
> Date: Mon, 9 Apr 2007 18:48:08 +0200
> 
> However, there is still another problem with these. I also wanted to set up 
> a keyboard macro for them. So I put
> 
> (fset 'anfein "«")
> (global-set-key "\C-z" 'anfein)
> 
> in my .emacs. This will work for all characters or strings, but does not for 
> « and ». I get an error message, "After 0 kbd macro iterations: Keyboard 
> macro terminated by a command ringing the bell" instead.

Works for me, if I evaluate these two lines in the *scratch* buffer.
If the same works for you, I suspect your .emacs's encoding might have
something to do with the problem.

Re: How to redefine Parantheses?

[Paul Batt]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 611 bytes --]

"Eli Zaretskii" <eliz@gnu.org> schrieb im Newsbeitrag 
news:mailman.1882.1176143496.7795.help-gnu-emacs@gnu.org...
>
> Works for me, if I evaluate these two lines in the *scratch* buffer.
> If the same works for you, I suspect your .emacs's encoding might have
> something to do with the problem.

those two lines evaluate well in the scratch buffer. But C-z does not work. 
Instead of inserting the «-character, I'm told "After 0 kbd macro 
iterations: Keyboard macro terminated by a command ringing the bell". I 
don't understand what that means. However, the macro works if I replace « 
with anything else. 

Re: How to redefine Parantheses?

[Ralf Angeli]

* Paul Batt (2007-04-10) writes:

> "Eli Zaretskii" <eliz@gnu.org> schrieb im Newsbeitrag 
> news:mailman.1882.1176143496.7795.help-gnu-emacs@gnu.org...
>>
>> Works for me, if I evaluate these two lines in the *scratch* buffer.
>> If the same works for you, I suspect your .emacs's encoding might have
>> something to do with the problem.
>
> those two lines evaluate well in the scratch buffer. But C-z does not work. 
> Instead of inserting the «-character, I'm told "After 0 kbd macro 
> iterations: Keyboard macro terminated by a command ringing the bell". I 
> don't understand what that means. However, the macro works if I replace « 
> with anything else. 

Your code works fine in CVS Emacs but it fails in Emacs 21.  Using
(defun anfein () (interactive) (insert "«"))
instead of your fset call should do the trick.

-- 
Ralf

Re: How to redefine Parantheses?

[Paul Batt]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 484 bytes --]

"Ralf Angeli" <dev.null@caeruleus.net> schrieb im Newsbeitrag 
news:461bcc86$0$6405$9b4e6d93@newsspool2.arcor-online.net...
>
> Your code works fine in CVS Emacs but it fails in Emacs 21.  Using
> (defun anfein () (interactive) (insert "«"))
> instead of your fset call should do the trick.

Your are right, thank you. Meanwhile I found the following to work also in 
Emacs 21:

(fset 'anfein
   [?\C-q ?2 ?5 ?3 return])

Emacs does that when you define the macro interactively.
Paul

Re: How to redefine Parantheses?

[Eli Zaretskii]

> From: "Paul Batt" <paul.batt@gmx.net>
> Date: Tue, 10 Apr 2007 08:36:10 +0200
> 
> "Eli Zaretskii" <eliz@gnu.org> schrieb im Newsbeitrag 
> news:mailman.1882.1176143496.7795.help-gnu-emacs@gnu.org...
> >
> > Works for me, if I evaluate these two lines in the *scratch* buffer.
> > If the same works for you, I suspect your .emacs's encoding might have
> > something to do with the problem.
> 
> those two lines evaluate well in the scratch buffer. But C-z does not work. 
> Instead of inserting the «-character, I'm told "After 0 kbd macro 
> iterations: Keyboard macro terminated by a command ringing the bell". I 
> don't understand what that means. However, the macro works if I replace « 
> with anything else. 

Very strange: as I said, it works for me.  So I guess, unless you or
someone else discovers what's wrong, exactly, we are stuck.

Hmm... maybe it has something to do with your ``language
environment''.  Can you show the text created by `report-emacs-bug',
where it states all the important environment variables and such
likes?

Re: How to redefine Parantheses?

[Paul Batt]

"Eli Zaretskii" <eliz@gnu.org> schrieb im Newsbeitrag 
news:mailman.1919.1176231393.7795.help-gnu-emacs@gnu.org...
> Very strange: as I said, it works for me.  So I guess, unless you or
> someone else discovers what's wrong, exactly, we are stuck.
>
> Hmm... maybe it has something to do with your ``language
> environment''.  Can you show the text created by `report-emacs-bug',
> where it states all the important environment variables and such
> likes?

Meanwhile I found that this works:

(fset 'anfein
   [?\C-q ?2 ?5 ?3 return])

Paul 

Re: How to redefine Parantheses?

[Peter Dyballa]

Am 08.04.2007 um 09:19 schrieb Paul Batt:

> I need German quotation marks « and » (ISO 8859 -1 Latin-1 171 and  
> 187).

These are French or Swiss ones. The German ones are „ and “.

The French are using them as in

	« start to end »

and the Swiss use them as in

	»from here to there«

--
Mit friedvollen Grüßen

   Pete

Remember: use logout to logout.