bug#52240: 29.0.50; Build failure after 35075267a6

bug#52240: 29.0.50; Build failure after 35075267a6

[Arash Esbati]

Hi all,

I can't build Emacs (Win10, Msys2) after the commit 35075267a6.  The
error is:

    Loading z:/pathto/emacs/lisp/international/characters.el (source)...
    Loading z:/pathto/emacs/lisp/international/charscript.el (source)...
    Loading z:/pathto/emacs/lisp/international/emoji-zwj.el (source)...
    Invalid read syntax: "\\N{left-to-right embedding}", 1, 0
    make[1]: *** [Makefile:908: bootstrap-emacs.pdmp] Error 127

Building with aaf0e62048 works.  This is with a fresh repo
(git clean -fdx).

Best, Arash

bug#52240: 29.0.50; Build failure after 35075267a6

[Lars Ingebrigtsen]

Arash Esbati <arash@gnu.org> writes:

>     Invalid read syntax: "\\N{left-to-right embedding}", 1, 0

Hm -- this works fine on Debian.  Does this mean that we can't use the
\N{...} things in source code, or does the w32 Emacs have to be set up
in a different way?

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#52240: 29.0.50; Build failure after 35075267a6

[Arash Esbati]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Arash Esbati <arash@gnu.org> writes:
>
>>     Invalid read syntax: "\\N{left-to-right embedding}", 1, 0
>
> Hm -- this works fine on Debian.  Does this mean that we can't use the
> \N{...} things in source code, or does the w32 Emacs have to be set up
> in a different way?

Hm, can't really tell, I'm not familiar the code in characters.el, but
this

    (insert ?\N{LATIN SMALL LETTER A WITH GRAVE})

inserts à.  Does this help?

bug#52240: 29.0.50; Build failure after 35075267a6

[Eli Zaretskii]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Date: Thu, 02 Dec 2021 11:21:08 +0100
> Cc: 52240@debbugs.gnu.org
> 
> Arash Esbati <arash@gnu.org> writes:
> 
> >     Invalid read syntax: "\\N{left-to-right embedding}", 1, 0
> 
> Hm -- this works fine on Debian.  Does this mean that we can't use the
> \N{...} things in source code, or does the w32 Emacs have to be set up
> in a different way?

It builds fine here, but I didn't try bootstrapping.

bug#52240: 29.0.50; Build failure after 35075267a6

[Lars Ingebrigtsen]

Arash Esbati <arash@gnu.org> writes:

> Hm, can't really tell, I'm not familiar the code in characters.el, but
> this
>
>     (insert ?\N{LATIN SMALL LETTER A WITH GRAVE})
>
> inserts à.  Does this help?

What about

(insert ?\N{latin small letter a with grave})

?  (Perhaps the identifier is case-sensitive on some systems...)

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#52240: 29.0.50; Build failure after 35075267a6

[Arash Esbati]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> What about
>
> (insert ?\N{latin small letter a with grave})
>
> ?  (Perhaps the identifier is case-sensitive on some systems...)

Works, inserts à.  Would be the first time where Windows is
case-sensitive 😉

bug#52240: 29.0.50; Build failure after 35075267a6

[Lars Ingebrigtsen]

Arash Esbati <arash@gnu.org> writes:

> Works, inserts à.  Would be the first time where Windows is
> case-sensitive 😉

What about what it's actually complaining about, then?  I.e., 

(insert ?\N{left-to-right embedding})

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#52240: 29.0.50; Build failure after 35075267a6

[Arash Esbati]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> What about what it's actually complaining about, then?  I.e., 
>
> (insert ?\N{left-to-right embedding})

Sorry, I thought I had tried that already but didn't look carefully at
the result.  When I do 'C-u C-x =' at the insertion, I get:

--8<---------------cut here---------------start------------->8---
             position: 308 of 308 (100%), column: 37
            character: ‪‬ (displayed as ‪‬) (codepoint 8234, #o20052, #x202a)
              charset: unicode (Unicode (ISO10646))
code point in charset: 0x202A
               script: symbol
               syntax: w 	which means: word
             category: L:Strong L2R
             to input: type "C-x 8 RET 202a" or "C-x 8 RET LEFT-TO-RIGHT EMBEDDING"
          buffer code: #xE2 #x80 #xAA
            file code: #xE2 #x80 #xAA (encoded by coding system utf-8-dos)
              display: by this font (glyph code):
    harfbuzz:-outline-Symbola-regular-normal-normal-serif-12-*-*-*-p-*-iso10646-1 (#x875)

Character code properties: customize what to show
  name: LEFT-TO-RIGHT EMBEDDING
  general-category: Cf (Other, Format)
  decomposition: (8234) ('‪')

There are text properties here:
  fontified            t
--8<---------------cut here---------------end--------------->8---

bug#52240: 29.0.50; Build failure after 35075267a6

[Eli Zaretskii]

> From: Arash Esbati <arash@gnu.org>
> Date: Thu, 02 Dec 2021 12:29:00 +0100
> Cc: 52240@debbugs.gnu.org
> 
> Lars Ingebrigtsen <larsi@gnus.org> writes:
> 
> > What about what it's actually complaining about, then?  I.e., 
> >
> > (insert ?\N{left-to-right embedding})
> 
> Sorry, I thought I had tried that already but didn't look carefully at
> the result.  When I do 'C-u C-x =' at the insertion, I get:
> 
> --8<---------------cut here---------------start------------->8---
>              position: 308 of 308 (100%), column: 37
>             character: ‪‬ (displayed as ‪‬) (codepoint 8234, #o20052, #x202a)
>               charset: unicode (Unicode (ISO10646))
> code point in charset: 0x202A
>                script: symbol
>                syntax: w 	which means: word
>              category: L:Strong L2R
>              to input: type "C-x 8 RET 202a" or "C-x 8 RET LEFT-TO-RIGHT EMBEDDING"
>           buffer code: #xE2 #x80 #xAA
>             file code: #xE2 #x80 #xAA (encoded by coding system utf-8-dos)
>               display: by this font (glyph code):
>     harfbuzz:-outline-Symbola-regular-normal-normal-serif-12-*-*-*-p-*-iso10646-1 (#x875)
> 
> Character code properties: customize what to show
>   name: LEFT-TO-RIGHT EMBEDDING
>   general-category: Cf (Other, Format)
>   decomposition: (8234) ('‪')

Which is correct, right?  or did you spot something problematic?

bug#52240: 29.0.50; Build failure after 35075267a6

[Lars Ingebrigtsen]

Arash Esbati <arash@gnu.org> writes:

> Lars Ingebrigtsen <larsi@gnus.org> writes:
>
>> What about what it's actually complaining about, then?  I.e., 
>>
>> (insert ?\N{left-to-right embedding})
>
> Sorry, I thought I had tried that already but didn't look carefully at
> the result.  When I do 'C-u C-x =' at the insertion, I get:

So that works too?  Then...  that build error is really odd.  Could it
be Emacs not loading the definitions of these things early enough on
some systems?

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#52240: 29.0.50; Build failure after 35075267a6

[Arash Esbati]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Arash Esbati <arash@gnu.org>
>> Date: Thu, 02 Dec 2021 12:29:00 +0100
>> Cc: 52240@debbugs.gnu.org
>> 
>> Lars Ingebrigtsen <larsi@gnus.org> writes:
>> 
>> > What about what it's actually complaining about, then?  I.e., 
>> >
>> > (insert ?\N{left-to-right embedding})
>> 
>> Sorry, I thought I had tried that already but didn't look carefully at
>> the result.  When I do 'C-u C-x =' at the insertion, I get:
>> 
>> --8<---------------cut here---------------start------------->8---
>>              position: 308 of 308 (100%), column: 37
>>             character: ‪‬ (displayed as ‪‬) (codepoint 8234, #o20052, #x202a)
>>               charset: unicode (Unicode (ISO10646))
>> code point in charset: 0x202A
>>                script: symbol
>>                syntax: w 	which means: word
>>              category: L:Strong L2R
>>              to input: type "C-x 8 RET 202a" or "C-x 8 RET LEFT-TO-RIGHT EMBEDDING"
>>           buffer code: #xE2 #x80 #xAA
>>             file code: #xE2 #x80 #xAA (encoded by coding system utf-8-dos)
>>               display: by this font (glyph code):
>>     harfbuzz:-outline-Symbola-regular-normal-normal-serif-12-*-*-*-p-*-iso10646-1 (#x875)
>> 
>> Character code properties: customize what to show
>>   name: LEFT-TO-RIGHT EMBEDDING
>>   general-category: Cf (Other, Format)
>>   decomposition: (8234) ('‪')
>
> Which is correct, right?  or did you spot something problematic?

This is correct.  So I eval'ed (C-x C-e) the example and the entries
in `glyphless--bidi-control-characters' :

--8<---------------cut here---------------start------------->8---
?\N{latin small letter a with grave}
224 (#o340, #xe0)

?\N{right-to-left embedding}
Debugger entered--Lisp error: (void-variable embedding})
  elisp--eval-last-sexp(nil)
  eval-last-sexp(nil)
  funcall-interactively(eval-last-sexp nil)
  command-execute(eval-last-sexp)

?\N{left-to-right override}
Debugger entered--Lisp error: (void-variable override})
  elisp--eval-last-sexp(nil)
  eval-last-sexp(nil)
  funcall-interactively(eval-last-sexp nil)
  command-execute(eval-last-sexp)

?\N{right-to-left override}
Debugger entered--Lisp error: (void-variable override})
  elisp--eval-last-sexp(nil)
  eval-last-sexp(nil)
  funcall-interactively(eval-last-sexp nil)
  command-execute(eval-last-sexp)

?\N{left-to-right isolate}
Debugger entered--Lisp error: (void-variable isolate})
  elisp--eval-last-sexp(nil)
  eval-last-sexp(nil)
  funcall-interactively(eval-last-sexp nil)
  command-execute(eval-last-sexp)

?\N{right-to-left isolate}
Debugger entered--Lisp error: (void-variable isolate})
  elisp--eval-last-sexp(nil)
  eval-last-sexp(nil)
  funcall-interactively(eval-last-sexp nil)
  command-execute(eval-last-sexp)

?\N{first strong isolate}
8296 (#o20150, #x2068)

?\N{pop directional formatting}
8236 (#o20054, #x202c)

?\N{pop directional isolate}
8297 (#o20151, #x2069)
--8<---------------cut here---------------end--------------->8---

It seems to me that something goes wrong when the unicode name contains
a hyphen?  As a random example,

?\N{LEFT-POINTING DOUBLE ANGLE QUOTATION MARK}

also complains about (void-variable MARK}).

bug#52240: 29.0.50; Build failure after 35075267a6

[Lars Ingebrigtsen]

Arash Esbati <arash@gnu.org> writes:

> This is correct.  So I eval'ed (C-x C-e) the example and the entries
> in `glyphless--bidi-control-characters' :

You can't `C-x C-e' these things -- it doesn't understand the syntax.
(There's a bug report for that.)  So you have to M-: and enter them
there (or put them inside a bigger sexp).

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#52240: 29.0.50; Build failure after 35075267a6

[Arash Esbati]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Arash Esbati <arash@gnu.org> writes:
>
>> This is correct.  So I eval'ed (C-x C-e) the example and the entries
>> in `glyphless--bidi-control-characters' :
>
> You can't `C-x C-e' these things -- it doesn't understand the syntax.
> (There's a bug report for that.)  So you have to M-: and enter them
> there (or put them inside a bigger sexp).

Ah, thanks, didn't know that.  They all work with M-:  Now I'm out of
ideas ☹️

bug#52240: 29.0.50; Build failure after 35075267a6

[Po Lu via Bug reports for GNU Emacs, the Swiss army knife of text editors]

Eli Zaretskii <eliz@gnu.org> writes:

>> Hm -- this works fine on Debian.  Does this mean that we can't use the
>> \N{...} things in source code, or does the w32 Emacs have to be set up
>> in a different way?

> It builds fine here, but I didn't try bootstrapping.

FWIW, I saw two people experience this issue, but bootstrapping made it
work for one of them.  One was running Arch "Linux", and the other was
running macOS.

bug#52240: 29.0.50; Build failure after 35075267a6

[Arash Esbati]

Po Lu <luangruo@yahoo.com> writes:

> Eli Zaretskii <eliz@gnu.org> writes:
>
>> It builds fine here, but I didn't try bootstrapping.
>
> FWIW, I saw two people experience this issue, but bootstrapping made it
> work for one of them.  One was running Arch "Linux", and the other was
> running macOS.

As a last resort, I tried:

git clone --depth=1 https://git.savannah.gnu.org/git/emacs.git
cd emacs
./autogen.sh
./configure
make

And the issue remains.

bug#52240: 29.0.50; Build failure after 35075267a6

[Glenn Morris]

Lars Ingebrigtsen wrote:

> Hm -- this works fine on Debian.

Fails for me as described, on Debian testing, RHEL 8.5, and also on
https://hydra.nixos.org/build/160128632

I note that international/emoji-zwj.el may not be fully regenerated in
some builds. Once again, I strongly recommend that in admin/unidata (and
similar), bootstrap-clean should run gen-clean. There was opposition
to this change previously.

bug#52240: 29.0.50; Build failure after 35075267a6

[Lars Ingebrigtsen]

Glenn Morris <rgm@gnu.org> writes:

> I note that international/emoji-zwj.el may not be fully regenerated in
> some builds. Once again, I strongly recommend that in admin/unidata (and
> similar), bootstrap-clean should run gen-clean. There was opposition
> to this change previously.

Ah!  Thanks; after doing an "extraclean", this fails for me, too.

Yes, I'm also in favour of having "make bootstrap" cleaning a bit more.
The added time to build stuff (which isn't massive) will probably save
us time in the long run (because we'd avoid these errors).

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#52240: 29.0.50; Build failure after 35075267a6

[Lars Ingebrigtsen]

Glenn Morris <rgm@gnu.org> writes:

> Fails for me as described, on Debian testing, RHEL 8.5, and also on
> https://hydra.nixos.org/build/160128632

This should now be fixed -- I made it use digits instead.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#52240: 29.0.50; Build failure after 35075267a6

[Arash Esbati]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Glenn Morris <rgm@gnu.org> writes:
>
>> Fails for me as described, on Debian testing, RHEL 8.5, and also on
>> https://hydra.nixos.org/build/160128632
>
> This should now be fixed -- I made it use digits instead.

Confirmed, Emacs builds again.  Thanks for fixing this.

Best, Arash