GNU bug report logs -
#21588
25.0.50: Single quotes becoming curvy quotes in message function
Previous Next
Reported by: Kaushal Modi <kaushal.modi <at> gmail.com>
Date: Tue, 29 Sep 2015 21:18:02 UTC
Severity: normal
Merged with 31597
Found in versions 25.0.50, 27.0.50
Done: Kaushal Modi <kaushal.modi <at> gmail.com>
Bug is archived. No further changes may be made.
To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 21588 in the body.
You can then email your comments to 21588 AT debbugs.gnu.org in the normal way.
Toggle the display of automated, internal messages from the tracker.
Report forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#21588; Package
emacs.
(Tue, 29 Sep 2015 21:18:02 GMT)
Full text and
rfc822 format available.
Acknowledgement sent
to
Kaushal Modi <kaushal.modi <at> gmail.com>:
New bug report received and forwarded. Copy sent to
bug-gnu-emacs <at> gnu.org.
(Tue, 29 Sep 2015 21:18:02 GMT)
Full text and
rfc822 format available.
Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
In an emacs -Q session, when you eval the below,
(message "'Hey'")
you get
’Hey’
"’Hey’"
Note the difference here:
This is what I get when I do "C-u C-x =" on the first single quote in the
message form:
position: 206 of 208 (99%), column: 14
character: ' (displayed as ') (codepoint 39, #o47, #x27)
preferred charset: ascii (ASCII (ISO646 IRV))
code point in charset: 0x27
script: latin
syntax: ' which means: prefix
category: .:Base, a:ASCII, l:Latin, r:Roman
to input: type "C-x 8 RET 27" or "C-x 8 RET APOSTROPHE"
buffer code: #x27
file code: #x27 (encoded by coding system utf-8-unix)
display: by this font (glyph code)
xft:-unknown-DejaVu Sans
Mono-normal-normal-normal-*-13-*-*-*-m-0-iso10646-1 (#x0A)
Character code properties: customize what to show
name: APOSTROPHE
old-name: APOSTROPHE-QUOTE
general-category: Po (Punctuation, Other)
decomposition: (39) (''')
There are text properties here:
face font-lock-string-face
fontified nil
This is what I get when I do "C-u C-x =" on the first single quote in the
*Messages* buffer (C-h e):
position: 67 of 338 (20%), column: 0
character: ’ (displayed as ’) (codepoint 8217, #o20031, #x2019)
preferred charset: unicode (Unicode (ISO10646))
code point in charset: 0x2019
script: symbol
syntax: . which means: punctuation
category: .:Base, c:Chinese, h:Korean, j:Japanese
to input: type "C-x 8 RET 2019" or "C-x 8 RET RIGHT SINGLE
QUOTATION MARK"
buffer code: #xE2 #x80 #x99
file code: #xE2 #x80 #x99 (encoded by coding system utf-8-unix)
display: by this font (glyph code)
xft:-unknown-DejaVu Sans
Mono-normal-normal-normal-*-13-*-*-*-m-0-iso10646-1 (#x6C1)
Character code properties: customize what to show
name: RIGHT SINGLE QUOTATION MARK
old-name: SINGLE COMMA QUOTATION MARK
general-category: Pf (Punctuation, Final quote)
decomposition: (8217) ('’')
I believe emacs should not be converting all APOSTROPHE characters to RIGHT
SINGLE QUOTATION MARK characters. The displayed messages look just plain
weird. To fix that, I would then need to replace all 'MSG' (apostrophe MSG
apostrophe) type message displays to `MSG' (back quote MSG apostrophe),
which is not practical.
Or, is there a defcustom I need to set to prevent this character
translation to curvy quotes?
Question:
Why is emacs converting all single
--
Kaushal Modi
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#21588; Package
emacs.
(Tue, 29 Sep 2015 21:33:02 GMT)
Full text and
rfc822 format available.
Message #8 received at 21588 <at> debbugs.gnu.org (full text, mbox):
This has been discussed to death and is documented in NEWS.
I'm not sure what else there could be left to say.
Reply sent
to
Kaushal Modi <kaushal.modi <at> gmail.com>:
You have taken responsibility.
(Tue, 29 Sep 2015 22:55:02 GMT)
Full text and
rfc822 format available.
Notification sent
to
Kaushal Modi <kaushal.modi <at> gmail.com>:
bug acknowledged by developer.
(Tue, 29 Sep 2015 22:55:03 GMT)
Full text and
rfc822 format available.
Message #13 received at 21588-done <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
Thanks for the reply.
Among all the noise about curved quotes, I missed out the discussion about
text-quoting-style. I started paying attention to that only recently when
Paul reverted curved quote stuff in the documentation.
I should have looked at NEWS; sorry about that.
I got my answer; I need to set text-quoting-style to 'straight.
For reference,
(progn
(message "text-quoting-style = default")
(let ((text-quoting-style))
(message "'Hey'")
(message "`Hey'"))
(message "text-quoting-style = curve")
(let ((text-quoting-style 'curve))
(message "'Hey'")
(message "`Hey'"))
(message "text-quoting-style = straight")
(let ((text-quoting-style 'straight))
(message "'Hey'")
(message "`Hey'"))
(message "text-quoting-style = grave")
(let ((text-quoting-style 'grave))
(message "'Hey'")
(message "`Hey'")))
gives
text-quoting-style = default
’Hey’
‘Hey’
text-quoting-style = curve
’Hey’
‘Hey’
text-quoting-style = straight
'Hey' [2 times]
text-quoting-style = grave
'Hey'
`Hey'
--
Kaushal Modi
On Tue, Sep 29, 2015 at 5:32 PM, Glenn Morris <rgm <at> gnu.org> wrote:
>
> This has been discussed to death and is documented in NEWS.
> I'm not sure what else there could be left to say.
>
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#21588; Package
emacs.
(Wed, 30 Sep 2015 01:06:02 GMT)
Full text and
rfc822 format available.
Message #16 received at 21588-done <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
Hi Glenn,
Can you please re-open this bug if the below argument makes sense?
It doesn't feel right that (message "'Hey'") gets displayed as ’Hey’. The
display looks "buggy" with the default value of text-quoting-style. A user
shouldn't be perplexed with that kind of display in an emacs -Q session. It
just feels wrong.
It would make more sense to have the default value of text-quoting-style as
'grave. That way, (message "'Hey'") prints 'Hey' and (message "`Hey'")
prints `Hey'; there are not surprises.
If we really want the curved quotes rendering, 'Hey' should NOT be rendered
as ’Hey’. It might be fine to render `Hey' as ‘Hey’. But having the default
setup render a common string like 'Hey' with identical curved quotes just
looks wrong.
Please review and reconsider based on the output given by the elisp snippet
in my previous email.
IMO, we are missing a "middle ground" option where if text-quoting-style is
set to 'middle-ground, (message "'Hey'") would yield 'Hey' (preserved
apostrophes) and (message "`Hey'") would give ‘Hey’ (balanced curved
quotes).
If the middle-ground option is decided to not be added, then fine. But
please don't have the unbalanced curved quotes ’Hey’ as the default.
I ended up opening this bug report after having quite a few emacs packages
print stuff with unbalanced curved quotes and that hit my OCD bar.
--
Kaushal Modi
On Tue, Sep 29, 2015 at 6:54 PM, Kaushal Modi <kaushal.modi <at> gmail.com>
wrote:
> Thanks for the reply.
>
> Among all the noise about curved quotes, I missed out the discussion about
> text-quoting-style. I started paying attention to that only recently when
> Paul reverted curved quote stuff in the documentation.
>
> I should have looked at NEWS; sorry about that.
>
> I got my answer; I need to set text-quoting-style to 'straight.
>
> For reference,
>
> (progn
> (message "text-quoting-style = default")
> (let ((text-quoting-style))
> (message "'Hey'")
> (message "`Hey'"))
> (message "text-quoting-style = curve")
> (let ((text-quoting-style 'curve))
> (message "'Hey'")
> (message "`Hey'"))
> (message "text-quoting-style = straight")
> (let ((text-quoting-style 'straight))
> (message "'Hey'")
> (message "`Hey'"))
> (message "text-quoting-style = grave")
> (let ((text-quoting-style 'grave))
> (message "'Hey'")
> (message "`Hey'")))
>
> gives
>
>
> text-quoting-style = default
> ’Hey’
> ‘Hey’
> text-quoting-style = curve
> ’Hey’
> ‘Hey’
> text-quoting-style = straight
> 'Hey' [2 times]
> text-quoting-style = grave
> 'Hey'
> `Hey'
>
>
>
>
>
>
> --
> Kaushal Modi
>
> On Tue, Sep 29, 2015 at 5:32 PM, Glenn Morris <rgm <at> gnu.org> wrote:
>
>>
>> This has been discussed to death and is documented in NEWS.
>> I'm not sure what else there could be left to say.
>>
>
>
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#21588; Package
emacs.
(Wed, 30 Sep 2015 21:12:01 GMT)
Full text and
rfc822 format available.
Message #19 received at 21588 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
Don't you mean "beautiful"? ;-)
That's the whole idea behind this, AFAICT.
The displayed messages look just plain weird.
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#21588; Package
emacs.
(Wed, 30 Sep 2015 22:41:01 GMT)
Full text and
rfc822 format available.
Message #22 received at 21588-done <at> debbugs.gnu.org (full text, mbox):
Kaushal Modi wrote:
> It doesn't feel right that (message "'Hey'") gets displayed as ’Hey’.
This shouldn't happen if your ~/.emacs has (setq text-quoting-style
'grave) or (setq text-quoting-style 'straight), so I assume the issue
came up because your text-quoting-style is ‘curve’ or defaults to ‘curve’.
The initial version of the ‘curve’ code behaved as you suggested, and
left isolated apostrophes alone. But as Dmitry pointed out, apostrophe
in English is represented by right single quotation mark as standard
typographic practice in fonts with directed quotes. In traditional
Emacs usage most isolated apostrophes in docstrings and diagnostics are
English apostrophes, so it seemed appropriate to default to
transliterating them as in English, of course with an escape hatch for
the exceptions intended to be straight apostrophes. So I changed the
code to translate all non-escaped apostrophes. This simplified the
transformation rules, which was also a win.
Although there were a few diagnostics in Elisp that quoted English
phrases 'like this', as I recall RMS said that directed quotes were
preferred, so I fixed the instances I found to quote `like this' in
message format strings. Undoubtedly I missed some instances and so some
minor display glitches remain, but they can be fixed as they turn up.
To get typewriter-style straight quoting in your own code, independent
of text-quoting-style, you can write (message "%s" "'Hey'").
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#21588; Package
emacs.
(Thu, 01 Oct 2015 03:55:02 GMT)
Full text and
rfc822 format available.
Message #25 received at 21588-done <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
> This shouldn't happen if your ~/.emacs has (setq text-quoting-style
'grave) or (setq text-quoting-style 'straight), so I assume the issue came
up because your text-quoting-style is ‘curve’ or defaults to ‘curve’.
Correct,
`text-quoting-style` was default (nil) and with that, (message "'Hey'") got
displayed as
’Hey’.
Now I need to set
`text-quoting-style` to 'grave so that apostrophes stay apostrophes and
grave quotes remain grave quotes; no surprising quote conversions. I would
vote for this to be the default.
> I missed some instances and so some minor display glitches remain, but
they can be fixed as they turn up.
Thanks, I am having difficulty grepping for an expression that has both
single and double quotes. But as I find such cases, I will report them or
submit a patch.
> To get typewriter-style straight quoting in your own code, independent of
text-quoting-style, you can write (message "%s" "'Hey'").
That is good to know, thank you!
Summary:
- Can the default be changed to 'grave?
- For the message forms that I or anyone else writes in future, is it
correct if I says that you must always use `Hey' format instead of 'Hey'
format? If so, we need to update the documentation that warns the user to
use the correct style.
Discussion:
As per "C-h i g (emacs) Quotation Marks" (excerpt below), we have
contradicting information.
One common way to quote is the typewriter convention, which quotes using
straight apostrophes 'like this' or double-quotes "like this". Another
common way is the curved quote convention, which uses left and right
single or double quotation marks ‘like this’ or “like this”. Typewriter
quotes are simple and portable; curved quotes are less ambiguous and
typically look nicer.
On one hand, the default text-quoting-style value of nil will render
(message 'like this') as
’like this’.
On the other hand, it is mentioned that typewriter quotes are simple and
portable.
I agree
that the typewriter
quotes
(straight quotes or apostrophes as I have been referring in my emails)
being simple and portable; the user does not need to think of whether to
enter ` or '. But then the default value of text-quoting-style as 'grave
makes more sense.
We would not want unwarned people using (message "'Hey'") in default
emacs sessions be faced with the
’Hey’
with unbalanced quotes.
If we want to keep the default of text-quoting-style as nil (or 'curve),
more logic needs to be added that decides whether left curly quote or right
curly quote should be used, which might have some performance impact as we
then need to check what the previous character is, what the next character
is, etc to decide if the current apostrophe needs to be replaced with a
left or a right quote.
Here
are
few common cases that I can think of
:
CASE 1: 'Hey' ->
‘Hey’
CASE 2: it's awesome ->
it
’
s awesome
CASE 3: my sisters' weddings ->
sisters
’
weddings
CASE 4: He said 'It's awesome!' ->
He said ‘It’s awesome!’ (It should have rather been -- He said "It's
awesome!" (use of double quotes). But I am just putting here all the
apostrophe cases I can think of..)
Based on above cases,
The apostrophe is replaced with left quote if,
(AND there's a non-space character after the apostrophe
(OR the apostrophe is first character in the line
there's white space before the apostrophe))
The apostrophe is replaced with right quote if,
there's a non-space character
before
the apostrophe
OR, probably..
If
there's a non-space character
before
the apostrophe, replace it with right quote, else use left quote.
I am by no means a literature or typography expert. So I am sure that the
above 4 cases are not enough to define the rules for left/right quotes. I
just wanted to emphasize that we need more intelligence in deciding which
quote to use if we want to keep the 'curly or nil as the default value of
text-quoting-style.
--
Kaushal Modi
[Message part 2 (text/html, inline)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#21588; Package
emacs.
(Fri, 02 Oct 2015 06:52:01 GMT)
Full text and
rfc822 format available.
Message #28 received at 21588-done <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
On 09/30/2015 08:54 PM, Kaushal Modi wrote:
> - Can the default be changed to 'grave?
The GNU coding standards were changed a while ago to recommend against grave
quoting in diagnostics and many GNU applications have already changed (e.g.,
coreutils, GCC). Although Emacs has been one of the holdouts, now is as good a
time as any to make the move.
> we need to update the documentation that warns the user to use the correct style.
Thanks, done in the attached patch, which I just installed in the master branch.
> it is mentioned that typewriter quotes are simple and portable.
That discussion was intended to be about text files, so I added a phrase to that
effect in the attached patch. This should help resolve the seeming contradiction.
> If we want to keep the default of text-quoting-style as nil (or 'curve), more logic needs to be added that decides whether left curly quote or right curly quote should be used, which might have some performance impact as we then need to check what the previous character is, what the next character is, etc to decide if the current apostrophe needs to be replaced with a left or a right quote.
It’s not a significant performance impact. It’s more a hassle of documenting a
more-complex approach. That is partly why I gave up on such an approach after
implementing it, and went with the simpler approach that is in Emacs master now.
> CASE 1: 'Hey' -> ‘Hey’
> CASE 2: it's awesome -> it’s awesome
> CASE 3: my sisters' weddings -> my sisters’ weddings
> CASE 4: He said 'It's awesome!' -> He said ‘It’s awesome!’
‘message’ already does cases 2 and 3 that way, as well as the 2nd apostrophe in
case 4. The remaining cases are quite rare in Emacs diagnostics, as the Emacs
style uses directed quotes when single-quoting.
> If there's a non-space character before the apostrophe, replace it with right quote, else use left quote.
That would mishandle common English phrases like “'60s pop culture” and “rock
'n' roll”. (These phrases don’t occur in Emacs diagnostics either, but as long
as we’re being pedantic....) Any heuristics will go awry sometimes so we need
some way to override the heuristics; and once we have such an override, there is
a real benefit to keeping the heuristics simple.
[0001-Fix-a-few-problems-with-directed-quotes.patch (text/plain, attachment)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#21588; Package
emacs.
(Fri, 02 Oct 2015 16:51:02 GMT)
Full text and
rfc822 format available.
Message #31 received at 21588-done <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
tl;dr: Thanks Paul, this is awesome!
---
> The GNU coding standards were changed a while ago to recommend against
grave quoting in diagnostics and many GNU applications have already changed
(e.g., coreutils, GCC). Although Emacs has been one of the holdouts, now
is as good a time as any to make the move.
I was unaware of that. Thanks for letting me know. For future reference for
anyone else, you can learn more about this "C-h i g (standards) Quote
Characters".
> Thanks, done in the attached patch, which I just installed in the master
branch.
Thanks! For future reference for anyone else, these clarifications can be
viewed in the following info nodes:
- (elisp) Displaying Messages
- (elisp) Signaling Errors
> That discussion was intended to be about text files, so I added a phrase
to that effect in the attached patch. This should help resolve the seeming
contradiction.
- (emacs) Quotation Marks
> Any heuristics will go awry sometimes so we need some way to override the
heuristics; and once we have such an override, there is a real benefit to
keeping the heuristics simple.
I agree to that argument.
Thanks for fixing the quotes in string formats in few other .el files.
In summary:
- I agree with you now that the current default value of text-quoting-style
is good.
- I will stick to that default and find places where the quoting needs to
be updated from 'Hey' to `Hey'.
- I will submit patches/pull requests wherever I find such instances within
emacs or in external packages.
Thank you for all your time and effort in making this clear and fixing up
the documentation and .el files.
--
Kaushal Modi
On Fri, Oct 2, 2015 at 2:51 AM, Paul Eggert <eggert <at> cs.ucla.edu> wrote:
> On 09/30/2015 08:54 PM, Kaushal Modi wrote:
> > - Can the default be changed to 'grave?
>
> The GNU coding standards were changed a while ago to recommend against
> grave quoting in diagnostics and many GNU applications have already changed
> (e.g., coreutils, GCC). Although Emacs has been one of the holdouts, now
> is as good a time as any to make the move.
>
> we need to update the documentation that warns the user to use the correct
>> style.
>>
>
> Thanks, done in the attached patch, which I just installed in the master
> branch.
>
> > it is mentioned that typewriter quotes are simple and portable.
>
> That discussion was intended to be about text files, so I added a phrase
> to that effect in the attached patch. This should help resolve the seeming
> contradiction.
>
> If we want to keep the default of text-quoting-style as nil (or 'curve),
>> more logic needs to be added that decides whether left curly quote or right
>> curly quote should be used, which might have some performance impact as we
>> then need to check what the previous character is, what the next character
>> is, etc to decide if the current apostrophe needs to be replaced with a
>> left or a right quote.
>>
>
> It’s not a significant performance impact. It’s more a hassle of
> documenting a more-complex approach. That is partly why I gave up on such
> an approach after implementing it, and went with the simpler approach that
> is in Emacs master now.
>
> > CASE 1: 'Hey' -> ‘Hey’
> > CASE 2: it's awesome -> it’s awesome
> > CASE 3: my sisters' weddings -> my sisters’ weddings
> > CASE 4: He said 'It's awesome!' -> He said ‘It’s awesome!’
>
> ‘message’ already does cases 2 and 3 that way, as well as the 2nd
> apostrophe in case 4. The remaining cases are quite rare in Emacs
> diagnostics, as the Emacs style uses directed quotes when single-quoting.
>
> If there's a non-space character before the apostrophe, replace it with
>> right quote, else use left quote.
>>
>
> That would mishandle common English phrases like “'60s pop culture” and
> “rock 'n' roll”. (These phrases don’t occur in Emacs diagnostics either,
> but as long as we’re being pedantic....) Any heuristics will go awry
> sometimes so we need some way to override the heuristics; and once we have
> such an override, there is a real benefit to keeping the heuristics simple.
>
[Message part 2 (text/html, inline)]
bug archived.
Request was from
Debbugs Internal Request <help-debbugs <at> gnu.org>
to
internal_control <at> debbugs.gnu.org.
(Sat, 31 Oct 2015 11:24:04 GMT)
Full text and
rfc822 format available.
bug unarchived.
Request was from
Glenn Morris <rgm <at> gnu.org>
to
control <at> debbugs.gnu.org.
(Sat, 26 May 2018 07:35:01 GMT)
Full text and
rfc822 format available.
Forcibly Merged 21588 31597.
Request was from
Glenn Morris <rgm <at> gnu.org>
to
control <at> debbugs.gnu.org.
(Sat, 26 May 2018 07:35:02 GMT)
Full text and
rfc822 format available.
bug archived.
Request was from
Debbugs Internal Request <help-debbugs <at> gnu.org>
to
internal_control <at> debbugs.gnu.org.
(Mon, 25 Jun 2018 11:24:06 GMT)
Full text and
rfc822 format available.
This bug report was last modified 5 years and 311 days ago.
Previous Next
GNU bug tracking system
Copyright (C) 1999 Darren O. Benham,
1997,2003 nCipher Corporation Ltd,
1994-97 Ian Jackson.