Non-ASCII characters in Info files produced by Texinfo 7.1

Non-ASCII characters in Info files produced by Texinfo 7.1

[Eli Zaretskii]

Texinfo 7.1, released a few days ago, changed the behavior wrt
characters produced from quotes, apostrophes and en- and em-dashes: by
default, it now produces the ASCII equivalents.  So, for example, "--"
in the Texinfo source remains "--" in the produced Info file, where
previous versions would produce a real en-dash '–'.

The main reason for this change is that it makes it easier to grep the
Info files.

To revert to previous behavior, we will need to pass an optional
argument to the makeinfo command in our Makefile's.  But before I make
this change, I would like to know if people here want the new behavior
or the old one.

For those who want the read about the reasons for this change in
behavior, here are the links:

  https://lists.gnu.org/archive/html/automake-patches/2022-12/msg00000.html
  https://lists.gnu.org/archive/html/automake-patches/2022-12/msg00019.html
  https://lists.gnu.org/archive/html/bug-texinfo/2023-06/msg00000.html

What do people think about this?

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Stefan Kangas]

Eli Zaretskii <eliz@gnu.org> writes:

> Texinfo 7.1, released a few days ago, changed the behavior wrt
> characters produced from quotes, apostrophes and en- and em-dashes: by
> default, it now produces the ASCII equivalents.  So, for example, "--"
> in the Texinfo source remains "--" in the produced Info file, where
> previous versions would produce a real en-dash '–'.

> To revert to previous behavior, we will need to pass an optional
> argument to the makeinfo command in our Makefile's.  But before I make
> this change, I would like to know if people here want the new behavior
> or the old one.

I think using "--" instead of "–" would be a step back typographically,
so whichever way we go, I think we should keep that.

IOW, if we want to go with the new behavior (fine by me), we should
s/--/–/ in doc/**/*texi.

> The main reason for this change is that it makes it easier to grep the
> Info files.

BTW, we could make the Emacs Info mode search for – when the user enters
--, for ‘ when the user enters `, and so on.  I think that would be a
useful feature.

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Visuwesh]

[திங்கள் அக்டோபர் 23, 2023] Stefan Kangas wrote:

> Eli Zaretskii <eliz@gnu.org> writes:
>
>> Texinfo 7.1, released a few days ago, changed the behavior wrt
>> characters produced from quotes, apostrophes and en- and em-dashes: by
>> default, it now produces the ASCII equivalents.  So, for example, "--"
>> in the Texinfo source remains "--" in the produced Info file, where
>> previous versions would produce a real en-dash '–'.
>
>> To revert to previous behavior, we will need to pass an optional
>> argument to the makeinfo command in our Makefile's.  But before I make
>> this change, I would like to know if people here want the new behavior
>> or the old one.
>
> I think using "--" instead of "–" would be a step back typographically,
> so whichever way we go, I think we should keep that.

+1, I like how the manuals are typeset currently.

> IOW, if we want to go with the new behavior (fine by me), we should
> s/--/–/ in doc/**/*texi.
>
>> The main reason for this change is that it makes it easier to grep the
>> Info files.
>
> BTW, we could make the Emacs Info mode search for – when the user enters
> --, for ‘ when the user enters `, and so on.  I think that would be a
> useful feature.

AFAIK, this is already true for `, see isearch-fold-quotes-mode.  We can
extend it to other similar replacements too.

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Stefan Kangas]

Visuwesh <visuweshm@gmail.com> writes:

> AFAIK, this is already true for `, see isearch-fold-quotes-mode.

Thanks, I didn't know that variable.  It should be enabled by default in
Info mode, and perhaps Help mode too.

> We can extend it to other similar replacements too.

That would be useful, indeed.

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Visuwesh]

[செவ்வாய் அக்டோபர் 24, 2023] Stefan Kangas wrote:

> Visuwesh <visuweshm@gmail.com> writes:
>
>> AFAIK, this is already true for `, see isearch-fold-quotes-mode.
>
> Thanks, I didn't know that variable.  It should be enabled by default in
> Info mode, and perhaps Help mode too.

It is enabled from Emacs 29 apparently:

    ** Isearch in "*Help*" and "*info*" now char-folds quote characters by default.
    This means that you can say 'C-s `foo' (GRAVE ACCENT) if the buffer
    contains "‘foo" (LEFT SINGLE QUOTATION MARK) and the like.  These
    quotation characters look somewhat similar in some fonts.  To switch
    this off, disable the new 'isearch-fold-quotes-mode' minor mode.

BTW, I forgot to mention this in my original message but there are cases
when the replacement of `' to their curly equivalent makes the text much
easier to read.  One particular case I can think of is:

   • Otherwise, if the command string contains ‘?’ surrounded by
     whitespace or ‘`?`’, Emacs runs the shell command once _for each
     file_, substituting the current file name for ‘?’ and ‘`?`’ each
     time.  You can use both ‘?’ and ‘`?`’ more than once in the
     command; the same file name replaces each occurrence.  If you mix
     them with ‘*’ the command signals an error.

in (info "(emacs) Shell Commands in Dired").  Although `?` is
highlighted here, the curly quotes make it easier to read IMO.  The same
with the docstring of dired-do-shell-command where no such highlighting
exists.

>> We can extend it to other similar replacements too.
>
> That would be useful, indeed.

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Eli Zaretskii]

> From: Visuwesh <visuweshm@gmail.com>
> Cc: Eli Zaretskii <eliz@gnu.org>,  emacs-devel@gnu.org
> Date: Tue, 24 Oct 2023 14:37:36 +0530
> 
> [செவ்வாய் அக்டோபர் 24, 2023] Stefan Kangas wrote:
> 
> > Visuwesh <visuweshm@gmail.com> writes:
> >
> >> AFAIK, this is already true for `, see isearch-fold-quotes-mode.
> >
> > Thanks, I didn't know that variable.  It should be enabled by default in
> > Info mode, and perhaps Help mode too.
> 
> It is enabled from Emacs 29 apparently:

It is enabled by default, but after some time in a long-running Emacs
session, it somehow gets turned off.  If someone can debug and fix
this, please do.

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Juri Linkov]

>> >> AFAIK, this is already true for `, see isearch-fold-quotes-mode.
>> >
>> > Thanks, I didn't know that variable.  It should be enabled by default in
>> > Info mode, and perhaps Help mode too.
>> 
>> It is enabled from Emacs 29 apparently:
>
> It is enabled by default, but after some time in a long-running Emacs
> session, it somehow gets turned off.  If someone can debug and fix
> this, please do.

It gets turned off as soon as you prefer to use other search type,
such as regexp/word search.  It can't match quotes in a regexp search.

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Eli Zaretskii]

> From: Juri Linkov <juri@linkov.net>
> Cc: Visuwesh <visuweshm@gmail.com>,  stefankangas@gmail.com,
>   emacs-devel@gnu.org
> Date: Wed, 25 Oct 2023 09:45:39 +0300
> 
> >> >> AFAIK, this is already true for `, see isearch-fold-quotes-mode.
> >> >
> >> > Thanks, I didn't know that variable.  It should be enabled by default in
> >> > Info mode, and perhaps Help mode too.
> >> 
> >> It is enabled from Emacs 29 apparently:
> >
> > It is enabled by default, but after some time in a long-running Emacs
> > session, it somehow gets turned off.  If someone can debug and fix
> > this, please do.
> 
> It gets turned off as soon as you prefer to use other search type,
> such as regexp/word search.  It can't match quotes in a regexp search.

Why isn't it get turned on when some Help command is invoked next?  I
use Help commands a lot, so if the above is the reason, the minor mode
should be back on upon the next C-h invocation.

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Eli Zaretskii]

> From: Juri Linkov <juri@linkov.net>
> Cc: Visuwesh <visuweshm@gmail.com>,  stefankangas@gmail.com,
>   emacs-devel@gnu.org
> Date: Wed, 25 Oct 2023 09:45:39 +0300
> 
> >> >> AFAIK, this is already true for `, see isearch-fold-quotes-mode.
> >> >
> >> > Thanks, I didn't know that variable.  It should be enabled by default in
> >> > Info mode, and perhaps Help mode too.
> >> 
> >> It is enabled from Emacs 29 apparently:
> >
> > It is enabled by default, but after some time in a long-running Emacs
> > session, it somehow gets turned off.  If someone can debug and fix
> > this, please do.
> 
> It gets turned off as soon as you prefer to use other search type,
> such as regexp/word search.  It can't match quotes in a regexp search.

I don't believe this is the case here: I almost never use other search
types, I barely even need to use Isearch in the *Help* buffers.  It
must be something else...

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Gerd Möllmann]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Juri Linkov <juri@linkov.net>
>> Cc: Visuwesh <visuweshm@gmail.com>,  stefankangas@gmail.com,
>>   emacs-devel@gnu.org
>> Date: Wed, 25 Oct 2023 09:45:39 +0300
>> 
>> >> >> AFAIK, this is already true for `, see isearch-fold-quotes-mode.
>> >> >
>> >> > Thanks, I didn't know that variable.  It should be enabled by default in
>> >> > Info mode, and perhaps Help mode too.
>> >> 
>> >> It is enabled from Emacs 29 apparently:
>> >
>> > It is enabled by default, but after some time in a long-running Emacs
>> > session, it somehow gets turned off.  If someone can debug and fix
>> > this, please do.
>> 
>> It gets turned off as soon as you prefer to use other search type,
>> such as regexp/word search.  It can't match quotes in a regexp search.
>
> I don't believe this is the case here: I almost never use other search
> types, I barely even need to use Isearch in the *Help* buffers.  It
> must be something else...

Maybe you could use add-variable-watcher?

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Alan Mackenzie]

Hello, Eli.

On Mon, Oct 23, 2023 at 22:12:21 +0300, Eli Zaretskii wrote:
> Texinfo 7.1, released a few days ago, changed the behavior wrt
> characters produced from quotes, apostrophes and en- and em-dashes: by
> default, it now produces the ASCII equivalents.  So, for example, "--"
> in the Texinfo source remains "--" in the produced Info file, where
> previous versions would produce a real en-dash '–'.

> The main reason for this change is that it makes it easier to grep the
> Info files.

> To revert to previous behavior, we will need to pass an optional
> argument to the makeinfo command in our Makefile's.  But before I make
> this change, I would like to know if people here want the new behavior
> or the old one.

Thanks for raising this issue on the mailing list.

> For those who want the read about the reasons for this change in
> behavior, here are the links:

>   https://lists.gnu.org/archive/html/automake-patches/2022-12/msg00000.html
>   https://lists.gnu.org/archive/html/automake-patches/2022-12/msg00019.html
>   https://lists.gnu.org/archive/html/bug-texinfo/2023-06/msg00000.html

> What do people think about this?

I have always disliked the replacement of ASCII characters by non-ASCII
Unicode punctuation in working documents; it makes those documents less
easy to search, whether by grep or by C-s, or whatever, and these
characters are displayed in an obtrusive face.

Quite likely, there will also be people who dislike the use of ASCII for
these purposes.

So I would vote for either texinfo's new default of using ASCII quotes,
etc., or having a configuration option to set/unset the command line
option to texinfo.

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[tomas]

[-- Attachment #1: Type: text/plain, Size: 325 bytes --]

On Mon, Oct 23, 2023 at 08:49:53PM +0000, Alan Mackenzie wrote:
> Hello, Eli.

[...]

> I have always disliked the replacement of ASCII characters by non-ASCII
> Unicode punctuation in working documents [...]

Not a devel here, by a long stretch, so count me as anecdotal, but I'm
with Alan here.

Cheers
-- 
t

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Christopher Dimech]

> Sent: Tuesday, October 24, 2023 at 4:31 PM
> From: tomas@tuxteam.de
> To: emacs-devel@gnu.org
> Subject: Re: Non-ASCII characters in Info files produced by Texinfo 7.1
>
> On Mon, Oct 23, 2023 at 08:49:53PM +0000, Alan Mackenzie wrote:
> > Hello, Eli.
>
> [...]
>
> > I have always disliked the replacement of ASCII characters by non-ASCII
> > Unicode punctuation in working documents [...]
>
> Not a devel here, by a long stretch, so count me as anecdotal, but I'm
> with Alan here.

One should learn a language and get an article published (e.g. in Japan).
It is evident that only American Isolationists dislike anything unresembling
the modern english alphabet.  Unicodes are your friends.  The only problem
people have is that their keyboard prints a different character when hitting
the punctuation marks.

A much more serious problem is Texinfo continuing to rely upon Plain Tex only.

> Cheers
> --
> t
>

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[tomas]

[-- Attachment #1: Type: text/plain, Size: 1335 bytes --]

On Tue, Oct 24, 2023 at 07:28:38AM +0200, Christopher Dimech wrote:
> > Sent: Tuesday, October 24, 2023 at 4:31 PM
> > From: tomas@tuxteam.de
> > To: emacs-devel@gnu.org
> > Subject: Re: Non-ASCII characters in Info files produced by Texinfo 7.1
> >
> > On Mon, Oct 23, 2023 at 08:49:53PM +0000, Alan Mackenzie wrote:
> > > Hello, Eli.
> >
> > [...]
> >
> > > I have always disliked the replacement of ASCII characters by non-ASCII
> > > Unicode punctuation in working documents [...]
> >
> > Not a devel here, by a long stretch, so count me as anecdotal, but I'm
> > with Alan here.
> 
> One should learn a language and get an article published (e.g. in Japan).
> It is evident that only American Isolationists dislike anything unresembling
> the modern english alphabet [...]

Please: keep your insinuations to yourself. I don't know whether Alan
is American or not, but AFAIK he's living somewhere else for quite
a while.

As for me, I've been to the USA (what you probably call "America") for
two weeks straight, that's all. Currently I'm more or less fluent on
four languages and have learnt two based on non-ASCII character sets.

> A much more serious problem is Texinfo continuing to rely upon Plain Tex only.

Now we're getting technical. Much better. I'd argue for LuaLaTex.

Cheers
-- 
t

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Yuri Khan]

On Tue, 24 Oct 2023 at 02:13, Eli Zaretskii <eliz@gnu.org> wrote:

> Texinfo 7.1, released a few days ago, changed the behavior wrt
> characters produced from quotes, apostrophes and en- and em-dashes: by
> default, it now produces the ASCII equivalents.  So, for example, "--"
> in the Texinfo source remains "--" in the produced Info file, where
> previous versions would produce a real en-dash '–'.
>
> The main reason for this change is that it makes it easier to grep the
> Info files.

The proper solution to that would be to (1) teach grep and generally
all the other search tools the notion of Unicode canonical and
compatibility equivalence, and/or (2) teach people to type proper
punctuation and to demand input devices that make typing proper
punctuation easier.

Using ' for open single quote/close single quote/single
prime/apostrophe, " for open double quote/close double quote/double
prime, hyphen(s) for minus/en dash/em dash was a technical limitation
of the typewriter. ` as a character separate from àèìòù was born as a
technical limitation of the typewriter. We are no longer using
typewriters and should not be clinging to their artifacts.

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Eli Zaretskii]

> From: Yuri Khan <yuri.v.khan@gmail.com>
> Date: Tue, 24 Oct 2023 13:54:26 +0700
> Cc: emacs-devel@gnu.org
> 
> Using ' for open single quote/close single quote/single
> prime/apostrophe, " for open double quote/close double quote/double
> prime, hyphen(s) for minus/en dash/em dash was a technical limitation
> of the typewriter.

That's not what the new default behavior of makeinfo does, AFAICT.  If
I'm missing something, please show examples of the behavior you
describe in Texinfo 7.1.

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Eli Zaretskii]

> Date: Tue, 24 Oct 2023 14:15:03 +0300
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: emacs-devel@gnu.org
> 
> > From: Yuri Khan <yuri.v.khan@gmail.com>
> > Date: Tue, 24 Oct 2023 13:54:26 +0700
> > Cc: emacs-devel@gnu.org
> > 
> > Using ' for open single quote/close single quote/single
> > prime/apostrophe, " for open double quote/close double quote/double
> > prime, hyphen(s) for minus/en dash/em dash was a technical limitation
> > of the typewriter.
> 
> That's not what the new default behavior of makeinfo does, AFAICT.

To clarify: I meant the "Using ' for open single quote/close single
quote" and " " for open double quote/close double quote" parts.  The
default quoting behavior was not changed in Texinfo 7.1.

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Manuel Giraud via Emacs development discussions.]

Eli Zaretskii <eliz@gnu.org> writes:

[...]

> What do people think about this?

I vote for keeping the old behaviour for "--" to "–" because, as Stefan
said, it is more typographically correct.

For ` and ', I think we should take the new behaviour.  It is less
unsettling than the en-dash and while a "C-s Don't" works out of the box
in Info-mode it does not work with "M-x occur" for instance.
-- 
Manuel Giraud

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Ulrich Mueller]

>>>>> On Tue, 24 Oct 2023, Manuel Giraud wrote:

> For ` and ', I think we should take the new behaviour.  It is less
> unsettling than the en-dash and while a "C-s Don't" works out of the box
> in Info-mode it does not work with "M-x occur" for instance.

IMHO the grave accent U+0060 looks horrible as a quotation mark,
especially with modern fonts.

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Manuel Giraud via Emacs development discussions.]

Ulrich Mueller <ulm@gentoo.org> writes:

>>>>>> On Tue, 24 Oct 2023, Manuel Giraud wrote:
>
>> For ` and ', I think we should take the new behaviour.  It is less
>> unsettling than the en-dash and while a "C-s Don't" works out of the box
>> in Info-mode it does not work with "M-x occur" for instance.
>
> IMHO the grave accent U+0060 looks horrible as a quotation mark,
> especially with modern fonts.

Yes, you're right.  It is not a problem for the font I'm using right now
but the ugliness is quite font-dependent.
-- 
Manuel Giraud

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Po Lu]

Ulrich Mueller <ulm@gentoo.org> writes:

>>>>>> On Tue, 24 Oct 2023, Manuel Giraud wrote:
>
>> For ` and ', I think we should take the new behaviour.  It is less
>> unsettling than the en-dash and while a "C-s Don't" works out of the box
>> in Info-mode it does not work with "M-x occur" for instance.
>
> IMHO the grave accent U+0060 looks horrible as a quotation mark,
> especially with modern fonts.

On the same note, Unicode quotation marks are displayed by a font
distinct from that used for ASCII text.  The upshot is that each line
with quotation marks is displayed several pixels taller than the rest,
producing a hideously incongruous mixture of lines varying in height.

To counteract this, I have configured the standard display table to
display Unicode quotation characters as their grave and apostrophe
counterparts, which I consider something Info mode ought to do by
default.  Yet others seem to know better.

But I digress; we should enable the Texinfo option that reverts to the
status quo, if not disable Unicode altogether, for the alternative of
replacing all dashes and quotation characters in texi source with
Unicode would render their editing agonizingly ungainly.

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Eli Zaretskii]

> From: Po Lu <luangruo@yahoo.com>
> Cc: Manuel Giraud via "Emacs development discussions."
>  <emacs-devel@gnu.org>,  Eli Zaretskii <eliz@gnu.org>,  Manuel Giraud
>  <manuel@ledu-giraud.fr>
> Date: Tue, 24 Oct 2023 18:59:35 +0800
> 
> Ulrich Mueller <ulm@gentoo.org> writes:
> 
> > IMHO the grave accent U+0060 looks horrible as a quotation mark,
> > especially with modern fonts.
> 
> On the same note, Unicode quotation marks are displayed by a font
> distinct from that used for ASCII text.

Not here, they aren't.  I suggest to use a different font for the
default face, if your current font doesn't handle these characters.

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Eli Zaretskii]

> From: Ulrich Mueller <ulm@gentoo.org>
> Cc: Eli Zaretskii <eliz@gnu.org>,  Manuel Giraud <manuel@ledu-giraud.fr>
> Date: Tue, 24 Oct 2023 10:21:11 +0200
> 
> >>>>> On Tue, 24 Oct 2023, Manuel Giraud wrote:
> 
> > For ` and ', I think we should take the new behaviour.  It is less
> > unsettling than the en-dash and while a "C-s Don't" works out of the box
> > in Info-mode it does not work with "M-x occur" for instance.
> 
> IMHO the grave accent U+0060 looks horrible as a quotation mark,
> especially with modern fonts.

Where do you see it used as quotation in the output of Texinfo 7.1?
Can you show an example?

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Eli Zaretskii]

> Cc: emacs-devel@gnu.org
> Date: Tue, 24 Oct 2023 10:02:46 +0200
> From:  Manuel Giraud via "Emacs development discussions." <emacs-devel@gnu.org>
> 
> For ` and ', I think we should take the new behaviour.

What is "the new behavior for ` and '" that you allude to?

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Manuel Giraud via Emacs development discussions.]

Eli Zaretskii <eliz@gnu.org> writes:

>> Cc: emacs-devel@gnu.org
>> Date: Tue, 24 Oct 2023 10:02:46 +0200
>> From:  Manuel Giraud via "Emacs development discussions." <emacs-devel@gnu.org>
>> 
>> For ` and ', I think we should take the new behaviour.
>
> What is "the new behavior for ` and '" that you allude to?

I meant: do not convert ` (GRAVE ACCENT) to ‘ (LEFT SINGLE QUOTATION
MARK) and do not convert ' (APOSTROPHE) to ’ (RIGHT SINGLE QUOTATION
MARK)
-- 
Manuel Giraud

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Eli Zaretskii]

> From: Manuel Giraud <manuel@ledu-giraud.fr>
> Cc: emacs-devel@gnu.org
> Date: Tue, 24 Oct 2023 14:28:30 +0200
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> >> Cc: emacs-devel@gnu.org
> >> Date: Tue, 24 Oct 2023 10:02:46 +0200
> >> From:  Manuel Giraud via "Emacs development discussions." <emacs-devel@gnu.org>
> >> 
> >> For ` and ', I think we should take the new behaviour.
> >
> > What is "the new behavior for ` and '" that you allude to?
> 
> I meant: do not convert ` (GRAVE ACCENT) to ‘ (LEFT SINGLE QUOTATION
> MARK) and do not convert ' (APOSTROPHE) to ’ (RIGHT SINGLE QUOTATION
> MARK)

AFAICT, the new default behavior does only the second part of this.

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Manuel Giraud via Emacs development discussions.]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Manuel Giraud <manuel@ledu-giraud.fr>
>> Cc: emacs-devel@gnu.org
>> Date: Tue, 24 Oct 2023 14:28:30 +0200
>> 
>> Eli Zaretskii <eliz@gnu.org> writes:
>> 
>> >> Cc: emacs-devel@gnu.org
>> >> Date: Tue, 24 Oct 2023 10:02:46 +0200
>> >> From:  Manuel Giraud via "Emacs development discussions." <emacs-devel@gnu.org>
>> >> 
>> >> For ` and ', I think we should take the new behaviour.
>> >
>> > What is "the new behavior for ` and '" that you allude to?
>> 
>> I meant: do not convert ` (GRAVE ACCENT) to ‘ (LEFT SINGLE QUOTATION
>> MARK) and do not convert ' (APOSTROPHE) to ’ (RIGHT SINGLE QUOTATION
>> MARK)
>
> AFAICT, the new default behavior does only the second part of this.

My bad.  From one of the links you post, I understood that it was for
both curly quotes.  Then I change my vote for the old behaviour
(i.e. produce LEFT and RIGHT SINGLE QUOTATION MARK) which, at least,
seems more consistent.
-- 
Manuel Giraud

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Eli Zaretskii]

> From: Manuel Giraud <manuel@ledu-giraud.fr>
> Cc: emacs-devel@gnu.org
> Date: Tue, 24 Oct 2023 15:43:12 +0200
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> >> >> For ` and ', I think we should take the new behaviour.
> >> >
> >> > What is "the new behavior for ` and '" that you allude to?
> >> 
> >> I meant: do not convert ` (GRAVE ACCENT) to ‘ (LEFT SINGLE QUOTATION
> >> MARK) and do not convert ' (APOSTROPHE) to ’ (RIGHT SINGLE QUOTATION
> >> MARK)
> >
> > AFAICT, the new default behavior does only the second part of this.
> 
> My bad.  From one of the links you post, I understood that it was for
> both curly quotes.  Then I change my vote for the old behaviour
> (i.e. produce LEFT and RIGHT SINGLE QUOTATION MARK) which, at least,
> seems more consistent.

Are there any other opinions?  For now, the opinions are 5:3 in favor
of keeping the pre-7.1 behavior of producing Unicode characters,
i.e. in favor of adding "-c ASCII_DASHES_AND_QUOTES=0" to our makeinfo
options.  If no other opinions will emerge within a few days that
change this picture, I will add that switch to our Makefile's.

Note that users can always countermand this by saying

  make INFO_OPTS='--no-split -c ASCII_DASHES_AND_QUOTES=1'

when they build Emacs.

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Alan Mackenzie]

Hello, Eli.

On Thu, Oct 26, 2023 at 11:53:39 +0300, Eli Zaretskii wrote:
> > From: Manuel Giraud <manuel@ledu-giraud.fr>
> > Cc: emacs-devel@gnu.org
> > Date: Tue, 24 Oct 2023 15:43:12 +0200

[ .... ]

> Are there any other opinions?  For now, the opinions are 5:3 in favor
> of keeping the pre-7.1 behavior of producing Unicode characters, i.e.
> in favor of adding "-c ASCII_DASHES_AND_QUOTES=0" to our makeinfo
> options.  If no other opinions will emerge within a few days that
> change this picture, I will add that switch to our Makefile's.

Please don't do that.  Please instead make it a configurable option,
something like --with-info-unicode-punctuation.

> Note that users can always countermand this by saying

>   make INFO_OPTS='--no-split -c ASCII_DASHES_AND_QUOTES=1'

> when they build Emacs.

"Always" is the word.  This means always editing the Makefile each time
it gets updated by a ./configure, or always having a constant
uncomfortable diff between the repository and one's working copy of
Makefile.in.

In practice, this will mean info will sometimes get built with the
Unicode punctuation, sometimes without.  The curse of this Unicode
punctuation in info will remain.  ;-)

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Eli Zaretskii]

> Date: Thu, 26 Oct 2023 09:43:37 +0000
> Cc: emacs-devel@gnu.org
> From: Alan Mackenzie <acm@muc.de>
> 
> > Are there any other opinions?  For now, the opinions are 5:3 in favor
> > of keeping the pre-7.1 behavior of producing Unicode characters, i.e.
> > in favor of adding "-c ASCII_DASHES_AND_QUOTES=0" to our makeinfo
> > options.  If no other opinions will emerge within a few days that
> > change this picture, I will add that switch to our Makefile's.
> 
> Please don't do that.  Please instead make it a configurable option,
> something like --with-info-unicode-punctuation.

When there's a 2:1 ratio in favor of the change, making a
configure-time option for it sounds like a lot of work without
justification.

> > Note that users can always countermand this by saying
> 
> >   make INFO_OPTS='--no-split -c ASCII_DASHES_AND_QUOTES=1'
> 
> > when they build Emacs.
> 
> "Always" is the word.  This means always editing the Makefile each time
> it gets updated by a ./configure, or always having a constant
> uncomfortable diff between the repository and one's working copy of
> Makefile.in.

No, I meant type the above from the command line.  No need to edit
anything.

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Alan Mackenzie]

Hello, Eli.

On Thu, Oct 26, 2023 at 13:11:46 +0300, Eli Zaretskii wrote:
> > Date: Thu, 26 Oct 2023 09:43:37 +0000
> > Cc: emacs-devel@gnu.org
> > From: Alan Mackenzie <acm@muc.de>

> > > Are there any other opinions?  For now, the opinions are 5:3 in favor
> > > of keeping the pre-7.1 behavior of producing Unicode characters, i.e.
> > > in favor of adding "-c ASCII_DASHES_AND_QUOTES=0" to our makeinfo
> > > options.  If no other opinions will emerge within a few days that
> > > change this picture, I will add that switch to our Makefile's.

> > Please don't do that.  Please instead make it a configurable option,
> > something like --with-info-unicode-punctuation.

> When there's a 2:1 ratio in favor of the change, ...

It's not 2:1, it was 5:3 on a very small sample size.  Now, with Gregory
chiming in, it appears to be 5:4.

> ... making a configure-time option for it sounds like a lot of work
> without justification.

I think there's justification.  How much work is it to introduce a new
config option?  I've never done it.

> > > Note that users can always countermand this by saying

> > >   make INFO_OPTS='--no-split -c ASCII_DASHES_AND_QUOTES=1'

> > > when they build Emacs.

> > "Always" is the word.  This means always editing the Makefile each time
> > it gets updated by a ./configure, or always having a constant
> > uncomfortable diff between the repository and one's working copy of
> > Makefile.in.

> No, I meant type the above from the command line.  No need to edit
> anything.

YUCK!  So after each make bootstrap, one will additionally have to go
through all the documentation directories, whether by hand or by script,
and redo the building of the info by a clumsy command line.

Again, how much work would it be to add such a configuration option?

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Eli Zaretskii]

> Date: Thu, 26 Oct 2023 10:55:13 +0000
> Cc: emacs-devel@gnu.org
> From: Alan Mackenzie <acm@muc.de>
> 
> > > >   make INFO_OPTS='--no-split -c ASCII_DASHES_AND_QUOTES=1'
> 
> > > > when they build Emacs.
> 
> > > "Always" is the word.  This means always editing the Makefile each time
> > > it gets updated by a ./configure, or always having a constant
> > > uncomfortable diff between the repository and one's working copy of
> > > Makefile.in.
> 
> > No, I meant type the above from the command line.  No need to edit
> > anything.
> 
> YUCK!  So after each make bootstrap, one will additionally have to go
> through all the documentation directories, whether by hand or by script,
> and redo the building of the info by a clumsy command line.

No, you say "make bootstrap INFO_OPTS='--no-split -c ASCII_DASHES_AND_QUOTES=1'.

> Again, how much work would it be to add such a configuration option?

More than zero, obviously.

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > When there's a 2:1 ratio in favor of the change,

How many people have actually stated a preference?
How have we contacted many users to ask what they prefer?

Before we count how many like A and how many like B, we had
better ask a large set of people -- poll the users.

-- 
Dr Richard Stallman (https://stallman.org)
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Ulrich Mueller]

>>>>> On Thu, 26 Oct 2023, Eli Zaretskii wrote:

> Are there any other opinions?  For now, the opinions are 5:3 in favor
> of keeping the pre-7.1 behavior of producing Unicode characters,
> i.e. in favor of adding "-c ASCII_DASHES_AND_QUOTES=0" to our makeinfo
> options.  If no other opinions will emerge within a few days that
> change this picture, I will add that switch to our Makefile's.

If most people think that the old behaviour is preferable, maybe also
ask the texinfo maintainers to change their default?

IMHO it doesn't make much sense if different parts of the GNU system
disagree from each other.

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Eli Zaretskii]

> From: Ulrich Mueller <ulm@gentoo.org>
> Cc: emacs-devel@gnu.org
> Date: Thu, 26 Oct 2023 11:54:15 +0200
> 
> >>>>> On Thu, 26 Oct 2023, Eli Zaretskii wrote:
> 
> > Are there any other opinions?  For now, the opinions are 5:3 in favor
> > of keeping the pre-7.1 behavior of producing Unicode characters,
> > i.e. in favor of adding "-c ASCII_DASHES_AND_QUOTES=0" to our makeinfo
> > options.  If no other opinions will emerge within a few days that
> > change this picture, I will add that switch to our Makefile's.
> 
> If most people think that the old behaviour is preferable, maybe also
> ask the texinfo maintainers to change their default?

I didn't yet decide the majority wants that.  That's why I'm
soliciting more opinions.

But that shouldn't preclude anyone from taking this up with the
Texinfo maintainers without waiting for the results of this poll, of
course.

Re: Non-ASCII characters in Info files produced by Texinfo 7.1

[Gregory Heytings]

>
> Are there any other opinions?  For now, the opinions are 5:3 in favor of 
> keeping the pre-7.1 behavior of producing Unicode characters
>

FWIW, I'm in favor of using the new behavior.