PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Nick Steeves]

[-- Attachment #1: Type: text/plain, Size: 155 bytes --]

Hi,

Here's the patch to fix the missing composited euro symbol issue I
mentioned earlier this week.  It's very simple, and has been tested.

Cheers,
Nick

[-- Attachment #2: emacs22-22.1.patch.gz --]
[-- Type: application/x-gzip, Size: 383 bytes --]

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Chong Yidong]

Nick Steeves <nick.steeves@uleth.ca> writes:

> The patch adds the euro symbol € to the auto-composed french input
> methods.  eg: typing e= changes the e= into €.  France uses the €, and
> it is also nice to be able to use it when writing emails.  The € is
> already found in the full latin input methods, and this patch can really
> be considered a trivial copy and paste from the latin section, to the
> french section.  I added the composing rule to all of the french
> methods, for completeness, although I only use french-postfix.

Can anyone who uses quail shed light on this proposed patch?  Would it
be of any use?


diff -ur ./leim/quail/latin-alt.el /tmp/emacs22-22.1/leim/quail/latin-alt.el
--- ./leim/quail/latin-alt.el	2007-01-16 01:40:03.000000000 -0700
+++ /tmp/emacs22-22.1/leim/quail/latin-alt.el	2008-02-26 16:39:47.000000000 -0700
@@ -1046,6 +1046,7 @@
  ("e'" ?é)
  ("e^" ?ê)
  ("e\"" ?ë)
+ ("e=" ?€)
  ("I^" ?Î)
  ("I\"" ?Ï)
  ("i^" ?î)
@@ -1075,6 +1076,7 @@
  ("e''" ["e'"])
  ("e^^" ["e^"])
  ("e\"\"" ["e\""])
+ ("e==" ["e="])
  ("I^^" ["I^"])
  ("I\"\"" ["I\""])
  ("i^^" ["i^"])
diff -ur ./leim/quail/latin-post.el /tmp/emacs22-22.1/leim/quail/latin-post.el
--- ./leim/quail/latin-post.el	2007-01-16 01:40:03.000000000 -0700
+++ /tmp/emacs22-22.1/leim/quail/latin-post.el	2008-02-26 16:41:15.000000000 -0700
@@ -1029,6 +1029,7 @@
  ("e'" ?é)
  ("e^" ?ê)
  ("e\"" ?ë)
+ ("e=" ?€)
  ("I^" ?Î)
  ("I\"" ?Ï)
  ("i^" ?î)
@@ -1058,6 +1059,7 @@
  ("e''" ["e'"])
  ("e^^" ["e^"])
  ("e\"\"" ["e\""])
+ ("e==" ["e=")
  ("I^^" ["I^"])
  ("I\"\"" ["I\""])
  ("i^^" ["i^"])
diff -ur ./leim/quail/latin-pre.el /tmp/emacs22-22.1/leim/quail/latin-pre.el
--- ./leim/quail/latin-pre.el	2007-01-16 01:40:03.000000000 -0700
+++ /tmp/emacs22-22.1/leim/quail/latin-pre.el	2008-02-26 16:42:29.000000000 -0700
@@ -307,6 +307,7 @@
  (",C" ?Ç)
  (",c" ?ç)
  (", " ?,)
+ ("~e" ?€)
 )
 
 (quail-define-package

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Miles Bader]

Chong Yidong <cyd@stupidchicken.com> writes:
> Can anyone who uses quail shed light on this proposed patch?  Would it
> be of any use?

It seems clearly useful -- people do, after all, use the euro in France!
The only issue, it would seem to me, is whether the proposed bindings
cause annoyance for users of those input methods.

Also, there should be a binding in the `C-x 8' keymap for the euro
symbol, though I don't know what it should be (maybe just `C-x 8 E',
which is similar to the bindings used for £ and ¥).

-Miles

-- 
Un-American, adj. Wicked, intolerable, heathenish.

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Stefan Monnier]

> Can anyone who uses quail shed light on this proposed patch?  Would it
> be of any use?

The latin-post version looks OK to me.  Intuitive and not
too intrusive, tho I'd use "E=" rather than "e=".

For the latin-pre version ("~e") I don't know what to think of it: it's
not terribly intuitive, and it is misplaced (see how the other entries
are ordred).  But at least it doesn't introduce a new prefix (like "=E"
would), so maybe the tilde is OK.
In anycase, I'd also use the uppercase E here, i.e. "~E" rather than "~e".


        Stefan

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Miles Bader]

Miles Bader <miles@gnu.org> writes:
> Also, there should be a binding in the `C-x 8' keymap for the euro
> symbol, though I don't know what it should be (maybe just `C-x 8 E',
> which is similar to the bindings used for £ and ¥).

Anyone object to adding this?

Thanks,

-Miles

-- 
Selfish, adj. Devoid of consideration for the selfishness of others.

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Stefan Monnier]

>> Also, there should be a binding in the `C-x 8' keymap for the euro
>> symbol, though I don't know what it should be (maybe just `C-x 8 E',
>> which is similar to the bindings used for £ and ¥).

> Anyone object to adding this?

I'd rather keep the space open for extensions,
so C-x 8 E = and/or C-x 8 = E would be preferable.


        Stefan

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Miles Bader]

Stefan Monnier <monnier@iro.umontreal.ca> writes:
>> Anyone object to adding this?
>
> I'd rather keep the space open for extensions,
> so C-x 8 E = and/or C-x 8 = E would be preferable.

Ok, but it seems rather inconsistent with the existing bindings...

-Miles

-- 
Guilt, n. The condition of one who is known to have committed an indiscretion,
as distinguished from the state of him who has covered his tracks.

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Juri Linkov]

> Stefan Monnier <monnier@iro.umontreal.ca> writes:
>>> Anyone object to adding this?
>>
>> I'd rather keep the space open for extensions,
>> so C-x 8 E = and/or C-x 8 = E would be preferable.
>
> Ok, but it seems rather inconsistent with the existing bindings...

Inconsistency would be too bad here, indeed.  Currently we have:

C-x 8 $         ¤
C-x 8 L         £
C-x 8 Y         ¥
C-x 8 c         ¢
C-x 8 * $       ¤
C-x 8 * L       £
C-x 8 * Y       ¥
C-x 8 * c       ¢

So either we should move them all to a new prefix, or just add `C-x 8 E'
as you suggested, and then some other new key could add a new free prefix
(like `C-x 8 =') later when necessary.

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Nick Steeves]

[-- Attachment #1: Type: text/plain, Size: 2618 bytes --]

Hi Stefan,

I chose ~e, because that is what the author of "latin-prefix" used,
though ~E makes more sense to me, as the actual key-sequence would then
be S-` S-e  As for the french and latin [alt-]postfix methods, is it ok
to keep it as e=, for simplicity, and to prevent an accidental S-e S-=
(which would result in E+ instead of E=)?  I'm not sure if someone is
going to write a new patch, but I fixed a few issues with the attached
patch.  My v1 patch had some issues...

Best regards,
Nick


On Tue, 2008-02-26 at 17:05 -0700, Nick Steeves wrote:
> Hi Handa, Chong,
> 
> Please note that I've attached a revised patch which puts the € rule in
> its proper alphabetized place in the stack of  quail rules, for the
> french-[alt-]postfix sections, and fixes a BUG I would have introduced
> in french-prefix.  It would have been a minor bug, but I hadn't noticed
> that the latin-prefix had already defined it's own method for composing
> the €.  I wrote "=e" instead of "~e", but that is fixed with the version
> 2 patch.  Also, I had made a small consistency mistake in latin-postfix,
> where I wrote ("e==" ?€) instead of ("e==" ]"e="]).  It didn't effect
> functionality for me, so I didn't notice anything when I tested the
> patch; I fixed this too.  The attached patch is far, far cleaner.  What
> is more, is it's usefully named.  Sorry about the generic useless
> filename of the last patch.
> 
> The patch adds the euro symbol € to the auto-composed french input
> methods.  eg: typing e= changes the e= into €.  France uses the €, and
> it is also nice to be able to use it when writing emails.  The € is
> already found in the full latin input methods, and this patch can really
> be considered a trivial copy and paste from the latin section, to the
> french section.  I added the composing rule to all of the french
> methods, for completeness, although I only use french-postfix.
> 
> Cheers,
> Nick


On Thu, 2008-02-28 at 16:31 -0500, Stefan Monnier wrote:
> > Can anyone who uses quail shed light on this proposed patch?  Would it
> > be of any use?
> 
> The latin-post version looks OK to me.  Intuitive and not
> too intrusive, tho I'd use "E=" rather than "e=".
> 
> For the latin-pre version ("~e") I don't know what to think of it: it's
> not terribly intuitive, and it is misplaced (see how the other entries
> are ordred).  But at least it doesn't introduce a new prefix (like "=E"
> would), so maybe the tilde is OK.
> In anycase, I'd also use the uppercase E here, i.e. "~E" rather than "~e".
> 
> 
>         Stefan


[-- Attachment #2: emacs22-22.1_add-euro-to-french-input-method_v2.patch.gz --]
[-- Type: application/x-gzip, Size: 438 bytes --]

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Eli Zaretskii]

> From: Miles Bader <miles@gnu.org>
> Date: Fri, 29 Feb 2008 05:00:00 +0900
> Cc: Kenichi Handa <handa@m17n.org>, Nick Steeves <nick.steeves@uleth.ca>,
> 	emacs-devel@gnu.org
> 
> Chong Yidong <cyd@stupidchicken.com> writes:
> > Can anyone who uses quail shed light on this proposed patch?  Would it
> > be of any use?
> 
> It seems clearly useful -- people do, after all, use the euro in France!

Yes, but the input methods that are suggested for modification are for
Latin-1, which AFAIK does not support the euro character.  Am I
missing something?

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Miles Bader]

Eli Zaretskii <eliz@gnu.org> writes:
>> > Can anyone who uses quail shed light on this proposed patch?  Would it
>> > be of any use?
>> 
>> It seems clearly useful -- people do, after all, use the euro in France!
>
> Yes, but the input methods that are suggested for modification are for
> Latin-1, which AFAIK does not support the euro character.  Am I
> missing something?

They're "french" and "italian" input methods; I don't see any inherent
connection with latin1 (at least the quail input-method definition
doesn't make any mention of coding systems).

[Indeed, if the patch submitter is the same guy I talked with on IRC, he
explicitly didn't want to use the generic latin-foo input methods
because they define too many useless bindings, he just wanted french
chars, plus euro.]

-Miles

-- 
Youth, n. The Period of Possibility, when Archimedes finds a fulcrum,
Cassandra has a following and seven cities compete for the honor of endowing a
living Homer.

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Juanma Barranquero]

On Fri, Feb 29, 2008 at 1:02 PM, Miles Bader <miles.bader@necel.com> wrote:

>  [Indeed, if the patch submitter is the same guy I talked with on IRC, he
>  explicitly didn't want to use the generic latin-foo input methods
>  because they define too many useless bindings, he just wanted french
>  chars, plus euro.]

It would be interesting to know which keyboard is he using. On a stock
Spanish keyboard, I get all spanish chars (accented vowels, ü, ñ, ç,
¡, ¿), and the euro sign it's at AltGr-e.

             Juanma

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Eli Zaretskii]

> From: Miles Bader <miles.bader@necel.com>
> Cc: cyd@stupidchicken.com, nick.steeves@uleth.ca, emacs-devel@gnu.org,
>         handa@m17n.org
> Date: Fri, 29 Feb 2008 21:02:07 +0900
> 
> > Yes, but the input methods that are suggested for modification are for
> > Latin-1, which AFAIK does not support the euro character.  Am I
> > missing something?
> 
> They're "french" and "italian" input methods; I don't see any inherent
> connection with latin1

Then what does the following excerpt from latin-1lt.el mean?

    (quail-define-package
     "italian-alt-postfix" "Latin-1" "IT<" t
 
According to the doc string of quail-define-package,

    quail-define-package is an autoloaded Lisp function in `quail'.
    (quail-define-package name language title &optional guidance docstring
    translation-keys forget-last-selection deterministic kbd-translate
    show-layout create-decode-map maximum-shortest overlay-plist
    update-translation-function conversion-keys simple)

    Define NAME as a new Quail package for input LANGUAGE.
    TITLE is a string to be displayed at mode-line to indicate this package.
 

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Miles Bader]

Eli Zaretskii <eliz@gnu.org> writes:
>> They're "french" and "italian" input methods; I don't see any inherent
>> connection with latin1
>
> Then what does the following excerpt from latin-1lt.el mean?
>
>     (quail-define-package
>      "italian-alt-postfix" "Latin-1" "IT<" t

Actually I was wrong about the input methods to which the patch applies
(I mis-counted line numbers):  it only changes various french input
methods which are all in the "French" language, not the "Latin-1"
language.

[To be honest, the Latin-1 "language" seems to be a pretty dubious idea
in the first place... perhaps that's something that should be revisited.]

-Miles

-- 
Christian, n. One who follows the teachings of Christ so long as they are not
inconsistent with a life of sin.

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Eli Zaretskii]

> From: Miles Bader <miles@gnu.org>
> Date: Fri, 29 Feb 2008 23:09:57 +0900
> Cc: cyd@stupidchicken.com, handa@m17n.org, nick.steeves@uleth.ca,
> 	emacs-devel@gnu.org
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> >> They're "french" and "italian" input methods; I don't see any inherent
> >> connection with latin1
> >
> > Then what does the following excerpt from latin-1lt.el mean?
> >
> >     (quail-define-package
> >      "italian-alt-postfix" "Latin-1" "IT<" t
> 
> Actually I was wrong about the input methods to which the patch applies
> (I mis-counted line numbers):  it only changes various french input
> methods which are all in the "French" language, not the "Latin-1"
> language.

Well, anyway: what encoding would users of latin-alt.el et al expect
for their files?  If they expect Latin-1, the euro will thwart that.

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Juanma Barranquero]

On Fri, Feb 29, 2008 at 3:20 PM, Eli Zaretskii <eliz@gnu.org> wrote:

>  Well, anyway: what encoding would users of latin-alt.el et al expect
>  for their files?

Latin-9, if they don't need the eight rarely-used chars removed from
Latin-1 to make room.

             Juanma

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Andreas Schwab]

Eli Zaretskii <eliz@gnu.org> writes:

> Well, anyway: what encoding would users of latin-alt.el et al expect
> for their files?  If they expect Latin-1, the euro will thwart that.

If they are using the Euro sign they probably expect Latin-9, unless
they use UTF-8 in the first place.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
PGP key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Andreas Schwab]

Eli Zaretskii <eliz@gnu.org> writes:

> Then what does the following excerpt from latin-1lt.el mean?
>
>     (quail-define-package
>      "italian-alt-postfix" "Latin-1" "IT<" t

That probably means there is a pasto, and "Latin-1" should have been
"Italian" in the first place.  And that probably applies to other input
methods as well.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
PGP key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Andreas Schwab]

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

>> Can anyone who uses quail shed light on this proposed patch?  Would it
>> be of any use?
>
> The latin-post version looks OK to me.  Intuitive and not
> too intrusive, tho I'd use "E=" rather than "e=".

latin-postfix already uses "e=", so this would be consistent.

> For the latin-pre version ("~e") I don't know what to think of it: it's
> not terribly intuitive, and it is misplaced (see how the other entries
> are ordred).  But at least it doesn't introduce a new prefix (like "=E"
> would), so maybe the tilde is OK.

Similar for latin-prefix and latin-9-prefix which already use "~e".

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
PGP key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Miles Bader]

Eli Zaretskii <eliz@gnu.org> writes:
>> Actually I was wrong about the input methods to which the patch applies
>> (I mis-counted line numbers):  it only changes various french input
>> methods which are all in the "French" language, not the "Latin-1"
>> language.
>
> Well, anyway: what encoding would users of latin-alt.el et al expect
> for their files?  If they expect Latin-1, the euro will thwart that.

Either latin-9 (which as I understand it, was more or less supposed to
replace latin-1...) or utf-8, presumably.  [In my experience utf-8 seems
to be mentioned most often these days.]

-Miles

-- 
Optimist, n. A proponent of the doctrine that black is white.

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Richard Stallman]

I think the main question is whether his choice of suffix
characters is the best choice.  Maybe it is.

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Richard Stallman]

    Yes, but the input methods that are suggested for modification are for
    Latin-1, which AFAIK does not support the euro character.  Am I
    missing something?

Is that the difference between Latin-1 and Latin-9?
I suspect that most of us use Latin-1 out of habit.

Those of us who prefer postfix have no choice.
There is a Latin-9-Prefix, but no Latin-9-Postfix.

If the solution is going to be "Stop using Latin-1 and use Latin-9",
we need to provide Latin-9 equivalents for all the Latin-1 input
methods.

Moreover, we should implement that advice to change to Latin-9 in the
country-specific input methods, such as French-Prefix and
French-Postfix.

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Jason Rumney]

Richard Stallman wrote:
> I think the main question is whether his choice of suffix
> characters is the best choice.  Maybe it is.
>   

They seem to be consistent with the few input methods that already have 
€ in them.

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[David Kastrup]

Richard Stallman <rms@gnu.org> writes:

>     Yes, but the input methods that are suggested for modification are for
>     Latin-1, which AFAIK does not support the euro character.  Am I
>     missing something?
>
> Is that the difference between Latin-1 and Latin-9?
> I suspect that most of us use Latin-1 out of habit.

It is the difference likely to be relevant for most Western people.
Latin-9 has no glyphs for 1/4, 1/2, 3/4, no isolated cedilla, acute and
diaeresis accents (which is not that terrible since there is no
well-defined composition method AFAIK), no "currency symbol" and
explicitly "broken bar".  What it has is Euro sign, S, Z, s and z with
caron (czech accent, I think), the French semi-characters oe and OE and
a capital Y with diaeresis.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Richard Stallman]

    If they are using the Euro sign they probably expect Latin-9, unless
    they use UTF-8 in the first place.

The choice of input method is separate from the choice of file coding
system.  In Emacs 22, there are different input methods for Latin-1
and Latin-9.  Even if you save as UTF-8, you have to choose one of them
or the other.

Is that different in Emacs 23?

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Richard Stallman]

    Latin-9, if they don't need the eight rarely-used chars removed from
    Latin-1 to make room.

What are these 8 characters?

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Juanma Barranquero]

On Sat, Mar 1, 2008 at 2:48 AM, Richard Stallman <rms@gnu.org> wrote:

>  What are these 8 characters?

The ones David explained:

¤ U+00A4 CURRENCY SIGN
¦ U+00A6 BROKEN BAR
¨ U+00A8 DIAERESIS
´ U+00B4 ACUTE ACCENT
¸ U+00B8 CEDILLA
¼ U+00BC VULGAR FRACTION ONE QUARTER
½ U+00BD VULGAR FRACTION ONE HALF
¾ U+00BE VULGAR FRACTION THREE QUARTERS

             Juanma

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Eli Zaretskii]

> From: Richard Stallman <rms@gnu.org>
> CC: miles@gnu.org, cyd@stupidchicken.com, nick.steeves@uleth.ca,
> 	emacs-devel@gnu.org, handa@m17n.org
> Date: Fri, 29 Feb 2008 14:54:31 -0500
> 
>     Yes, but the input methods that are suggested for modification are for
>     Latin-1, which AFAIK does not support the euro character.  Am I
>     missing something?
> 
> Is that the difference between Latin-1 and Latin-9?
> I suspect that most of us use Latin-1 out of habit.
> 
> Those of us who prefer postfix have no choice.
> There is a Latin-9-Prefix, but no Latin-9-Postfix.
> 
> If the solution is going to be "Stop using Latin-1 and use Latin-9",
> we need to provide Latin-9 equivalents for all the Latin-1 input
> methods.

What bothers me is that the same input method suddenly switches the
encoding I need to use from under my feet, just because I used one
particular character supported by that input method.  And on top of
that, if I already used some of the characters not in Latin-9, my only
practical choice is UTF-8.  Somehow, this sounds wrong to me, but it
looks like I'm the only one bothered by that, so maybe this problem
exists only in my eyes.

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[David Kastrup]

Eli Zaretskii <eliz@gnu.org> writes:

> What bothers me is that the same input method suddenly switches the
> encoding I need to use from under my feet, just because I used one
> particular character supported by that input method.

Huh?  That's not related to the input method at all.  The input method
switches nothing.  It is the presence of the character (and in the
unicode-2 branch, we don't even have to unify those characters anymore)
in the buffer that requires some specific coding.

> And on top of that, if I already used some of the characters not in
> Latin-9, my only practical choice is UTF-8.  Somehow, this sounds
> wrong to me, but it looks like I'm the only one bothered by that, so
> maybe this problem exists only in my eyes.

It is not related to input methods.  In fact, I find it somewhat silly
to have different latin-x input methods at all.  Why don't we replace
them all with a single latin-x input method, or with language specific
ones?  It seems somewhat pointless to have input methods that
meticulously avoid producing a certain character when this character can
be produced by any number of other means.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Kenichi Handa]

In article <8563w6vipk.fsf@lola.goethe.zz>, David Kastrup <dak@gnu.org> writes:

> Eli Zaretskii <eliz@gnu.org> writes:
> > What bothers me is that the same input method suddenly switches the
> > encoding I need to use from under my feet, just because I used one
> > particular character supported by that input method.

> Huh?  That's not related to the input method at all.  The input method
> switches nothing.  It is the presence of the character (and in the
> unicode-2 branch, we don't even have to unify those characters anymore)
> in the buffer that requires some specific coding.

> > And on top of that, if I already used some of the characters not in
> > Latin-9, my only practical choice is UTF-8.  Somehow, this sounds
> > wrong to me, but it looks like I'm the only one bothered by that, so
> > maybe this problem exists only in my eyes.

> It is not related to input methods.  In fact, I find it somewhat silly
> to have different latin-x input methods at all.  Why don't we replace
> them all with a single latin-x input method, or with language specific
> ones?  It seems somewhat pointless to have input methods that
> meticulously avoid producing a certain character when this character can
> be produced by any number of other means.

I can think of one reason to have them.  While one is using
latin-x-* input methods, he can think that he never type a
character not included in latin-x.  In other word, if he
wants to type a character not supproted by latin-x-* input
methods, he must expect that his file can't be saved by
latin-x.

---
Kenichi Handa
handa@ni.aist.go.jp

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[David Kastrup]

Kenichi Handa <handa@m17n.org> writes:

> In article <8563w6vipk.fsf@lola.goethe.zz>, David Kastrup <dak@gnu.org> writes:
>> In fact, I find it somewhat silly to have different latin-x input
>> methods at all.  Why don't we replace them all with a single latin
>> input method, or with language specific ones?  It seems somewhat
>> pointless to have input methods that meticulously avoid producing a
>> certain character when this character can be produced by any number
>> of other means.
>
> I can think of one reason to have them.  While one is using
> latin-x-* input methods, he can think that he never type a
> character not included in latin-x.  In other word, if he
> wants to type a character not supproted by latin-x-* input
> methods, he must expect that his file can't be saved by
> latin-x.

So he does when using cut&paste, or quite a few other means of producing
characters.  I really don't think that we are doing people much of a
favor of maintaining a lot of slightly different character-set specific
input methods.  It would make more sense to have some mechanism complain
upon any way of inserting non-encodable characters (using buffer
modification hooks, for example), and then have just one generic
well-maintained latin input method rather than a dozen "forks".  It
would also be possible to auto-prune input methods depending on the
desired target encoding (for example, use the TeX input encoding for
accented letters but tell it not to produce things outside of latin-9).

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Eli Zaretskii]

> From: David Kastrup <dak@gnu.org>
> Cc: rms@gnu.org,  cyd@stupidchicken.com,  nick.steeves@uleth.ca,  emacs-devel@gnu.org,  handa@m17n.org,  miles@gnu.org
> Date: Sat, 01 Mar 2008 11:31:35 +0100
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > What bothers me is that the same input method suddenly switches the
> > encoding I need to use from under my feet, just because I used one
> > particular character supported by that input method.
> 
> Huh?  That's not related to the input method at all.  The input method
> switches nothing.  It is the presence of the character (and in the
> unicode-2 branch, we don't even have to unify those characters anymore)
> in the buffer that requires some specific coding.

Sigh.  You could have assumed that I knew all that, couldn't you?  I
said it switches the encoding _I_need_to_use_ (well, "changes" would
be a better word than "switches"), meaning the encoding I will have to
use when saving the buffer.

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Eli Zaretskii]

> From: David Kastrup <dak@gnu.org>
> Cc: eliz@gnu.org,  rms@gnu.org,  cyd@stupidchicken.com,  emacs-devel@gnu.org,  nick.steeves@uleth.ca,  miles@gnu.org
> Date: Sat, 01 Mar 2008 12:51:29 +0100
> 
> > I can think of one reason to have them.  While one is using
> > latin-x-* input methods, he can think that he never type a
> > character not included in latin-x.  In other word, if he
> > wants to type a character not supproted by latin-x-* input
> > methods, he must expect that his file can't be saved by
> > latin-x.
> 
> So he does when using cut&paste, or quite a few other means of producing
> characters.

That's true, but it doesn't make the problem easier.

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[David Kastrup]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: David Kastrup <dak@gnu.org>
>> Cc: eliz@gnu.org,  rms@gnu.org,  cyd@stupidchicken.com,  emacs-devel@gnu.org,  nick.steeves@uleth.ca,  miles@gnu.org
>> Date: Sat, 01 Mar 2008 12:51:29 +0100
>> 
>> > I can think of one reason to have them.  While one is using
>> > latin-x-* input methods, he can think that he never type a
>> > character not included in latin-x.  In other word, if he
>> > wants to type a character not supproted by latin-x-* input
>> > methods, he must expect that his file can't be saved by
>> > latin-x.
>> 
>> So he does when using cut&paste, or quite a few other means of producing
>> characters.
>
> That's true, but it doesn't make the problem easier.

The point was that input methods may not be the best place to tackle
this problem (which does not happen to be a problem for me, but that's a
different question).

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Andreas Schwab]

Richard Stallman <rms@gnu.org> writes:

>     If they are using the Euro sign they probably expect Latin-9, unless
>     they use UTF-8 in the first place.
>
> The choice of input method is separate from the choice of file coding
> system.  In Emacs 22, there are different input methods for Latin-1
> and Latin-9.  Even if you save as UTF-8, you have to choose one of them
> or the other.

I agree with David: a "Latin-1" or "Latin-9" input method does not make
any sense any more, now that everything is unified.  The main reason for
different input methods that should remain is to represent often used
characters in a particular language and make them easy to type.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
PGP key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Juanma Barranquero]

On Sat, Mar 1, 2008 at 9:57 AM, Eli Zaretskii <eliz@gnu.org> wrote:

>  What bothers me is that the same input method suddenly switches the
>  encoding I need to use from under my feet, just because I used one
>  particular character supported by that input method.  And on top of
>  that, if I already used some of the characters not in Latin-9, my only
>  practical choice is UTF-8.

It boils down to what do you think it is more likely:

  a) The user types characters outside latin-1 (or whatever) because
he needs them, and will have to worry later how to save the buffer
  b) The user wants to save in latin-1, uses characters outside it by
accident (because they are available in the input method), and will
have to worry later about removing or substituting them.

             Juanma

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Eli Zaretskii]

> From: David Kastrup <dak@gnu.org>
> Cc: handa@m17n.org,  rms@gnu.org,  cyd@stupidchicken.com,  emacs-devel@gnu.org,  nick.steeves@uleth.ca,  miles@gnu.org
> Date: Sat, 01 Mar 2008 14:00:14 +0100
> 
> >> So he does when using cut&paste, or quite a few other means of producing
> >> characters.
> >
> > That's true, but it doesn't make the problem easier.
> 
> The point was that input methods may not be the best place to tackle
> this problem

I agree.

All I was worried about was whether users do assume the current
latin-* input methods do not produce characters outside a single Latin
charset.  If users from the European cultures don't mind, neither do
I.

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Jason Rumney]

Eli Zaretskii wrote:
> All I was worried about was whether users do assume the current
> latin-* input methods do not produce characters outside a single Latin
> charset.  If users from the European cultures don't mind, neither do
> I.
>   

If the input method is called "latin-1-prefix" (-postfix, -alt), then 
the user would expect it not to produce characters outside the latin-1 
range. But for input methods called "french-prefix" etc, I think they 
would expect to be able to type all letters and symbols in common usage 
in French, which these days includes the euro symbol. European users 
have surely encountered the fact that the euro is not covered by the 
latin-1 character set by now, so if they want to save in latin-1 
encoding, they will simply avoid it. The "latin-*" input methods already 
contain the euro, so there is precedent for including it.

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Richard Stallman]

    It is not related to input methods.  In fact, I find it somewhat silly
    to have different latin-x input methods at all.  Why don't we replace
    them all with a single latin-x input method, or with language specific
    ones?

We should try to write general latin-everything input methods, but
since they will have lots of combinations to make for each letter,
they may be quite inconvenient to USE.  There may be too many active
prefixes or suffixes, and they can get in the way of ordinary typing.

Making the general latin-everything as convenient as is feasible
be could be a difficult job of tuning.

Until we see that people prefer these general methods to the
individual Latin-N input methods, we should not even think of deleting
those.

      It would make more sense to have some mechanism complain
    upon any way of inserting non-encodable characters (using buffer
    modification hooks, for example), and then have just one generic
    well-maintained latin input method rather than a dozen "forks".

Trying to save with characters that cannot be encoded will already
warn you.  I am not sure it is useful to detect those characters when
they are inserted, but if it is, there are various ways to do that.
Maybe we could make Font Lock highlight them.

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Stefan Monnier]

> If the input method is called "latin-1-prefix" (-postfix, -alt), then the
> user would expect it not to produce characters outside the latin-1 
> range. But for input methods called "french-prefix" etc, I think they would
> expect to be able to type all letters and symbols in common usage in French,
> which these days includes the euro symbol. European users have surely
> encountered the fact that the euro is not covered by the latin-1 character
> set by now, so if they want to save in latin-1 
> encoding, they will simply avoid it. The "latin-*" input methods already
> contain the euro, so there is precedent for including it.

Most of the latin-N input methods made a lot of sense in Emacs-22 where
the chars were not unified.  Now that they're unified, the distinction
is not nearly as important, so I'm in favor of merging/removing most of
those methods, keeping just some generic "latin" input method maybe,
plus the language-specific ones of course.

The impact it may have on the encoding used when saving the file is best
tackled elsewhere.  E.g. if the user's preferred coding systems
indicate that latin-1 is preferred over latin-9, then saving a file as
latin-9 in the absence of any coding cookie should prompt the user,
explaining that Emacs will mis-detect it as latin-1 next time around.


        Stefan

Re: PATCH to add euro symbol to French-{postfix, alt-postfix, prefix} input methods

[Stefan Monnier]

> latin-postfix already uses "e=", so this would be consistent.
> Similar for latin-prefix and latin-9-prefix which already use "~e".

I guess that's OK then.


        Stefan