bug#43822: 28.0.50; build failed on revno.5ec2115

bug#43822: 28.0.50; build failed on revno.5ec2115

[Katsumi Yamaoka]

Hi,

On Sun, 04 Oct 2020 15:41:59 -0400, Juri Linkov wrote:
> branch: master
> commit 5ec21155c39aab8a452d190a260e6912d1d9a920
> Author: Juri Linkov <juri@linkov.net>
> Commit: Juri Linkov <juri@linkov.net>

>     Use '…' for ellipsis in truncate-string-to-width by default (bug#41250)

Probably because of this change, building Emacs got to fail (on
at least Cygwin platform).  Before the change (96a8e8) it's ok.
An output of make is below.  Where I added (setq debug-on-error t)
to `skkdic-convert-postfix' in "lisp/international/ja-dic-cnv.el".

Note that Emacs tries to load the file named "...".  I couldn't
find how `lookup-nested-alist' calls `load-with-code-conversion',
sorry.

[...]
make[2]: Entering directory '/Work/Emacs/lisp'
make -C ../leim all EMACS="../src/bootstrap-emacs.exe"
make[3]: Entering directory '/Work/Emacs/leim'
  GEN      ../lisp/leim/ja-dic/ja-dic.el
  INFO     Processing OKURI-ARI entries
  INFO     Processing POSTFIX entries
Debugger entered--Lisp error: (invalid-read-syntax "?")
  read(#<buffer  *load*>)
  eval-buffer(#<buffer  *load*> nil "/Work/Emacs/lisp/international/..." nil t)  ; Reading at buffer position 2307
[...]
  load-with-code-conversion("/Work/Emacs/lisp/international/..." "/Work/Emacs/lisp/international/..." nil t)
  lookup-nested-alist("行" (skkdic-postfix-list) nil nil t)
  skkdic-convert-postfix(#<buffer  *skkdic-unannotated*> #<buffer *skkdic-work*>)
  skkdic-convert("/Work/Emacs/leim/SKK-DIC/SKK-JI..." "/Work/Emacs/lisp/leim/ja-dic")
  batch-skkdic-convert()
  funcall(batch-skkdic-convert)
[...]
make[3]: *** [Makefile:143: ../lisp/leim/ja-dic/ja-dic.el] Error 255
make[3]: Leaving directory '/Work/Emacs/leim'
make[2]: *** [Makefile:352: leim] Error 2
make[2]: Leaving directory '/Work/Emacs/lisp'
make[1]: *** [Makefile:794: ../lisp/loaddefs.el] Error 2
make[1]: Leaving directory '/Work/Emacs/src'
make: *** [Makefile:424: src] Error 2

Thanks.

Re: master 5ec2115: Use '…' for ellipsis in truncate-string-to-width by default (bug#41250)

[Herbert J. Skuhra]

On Sun, 04 Oct 2020 21:41:59 +0200, Juri Linkov wrote:
> 
> branch: master
> commit 5ec21155c39aab8a452d190a260e6912d1d9a920
> Author: Juri Linkov <juri@linkov.net>
> Commit: Juri Linkov <juri@linkov.net>
> 
>     Use '…' for ellipsis in truncate-string-to-width by default (bug#41250)
>     
>     * lisp/international/mule-util.el (truncate-string-ellipsis):
>     Change the default value to nil.
>     (truncate-string-ellipsis): New function.
>     (truncate-string-to-width): Use the value returned from the
>     function 'truncate-string-ellipsis'.
>     
>     * lisp/tab-bar.el (tab-bar-tab-name-truncated):
>     * lisp/tab-line.el (tab-line-tab-name-ellipsis):
>     Take advantage of the improvement of the ellipsis default value
>     in truncate-string-to-width and truncate-string-ellipsis.
>     
>     * doc/lispref/display.texi (Size of Displayed Text):
>     Improve description of truncate-string-ellipsis.
> ---
>  doc/lispref/display.texi        |  3 ++-
>  etc/NEWS                        |  6 ++++++
>  lisp/international/mule-util.el | 21 +++++++++++++++++----
>  lisp/tab-bar.el                 | 10 +++-------
>  lisp/tab-line.el                |  3 +--
>  5 files changed, 29 insertions(+), 14 deletions(-)

Hi,

after this commit I can no longer do a clean build:

% git clean -xfd
% ./autogen.sh
% ./configure --prefix=/usr/local/emacs --with-x-toolkit=lucid --with-mailutils
% make

  GEN      ../lisp/leim/ja-dic/ja-dic.el
  INFO     Processing OKURI-ARI entries
  INFO     Processing POSTFIX entries
Loading macroexp.elc...
Invalid read syntax: "?"
make[3]: *** [Makefile:143: ../lisp/leim/ja-dic/ja-dic.el] Error 255
make[3]: Leaving directory '/home/herbert/source/emacs/leim'
make[2]: *** [Makefile:352: leim] Error 2
make[2]: Leaving directory '/home/herbert/source/emacs/lisp'
make[1]: *** [Makefile:794: ../lisp/loaddefs.el] Error 2
make[1]: Leaving directory '/home/herbert/source/emacs/src'
make: *** [Makefile:424: src] Error 2

--
Herbert

Re: master 5ec2115: Use '…' for ellipsis in truncate-string-to-width by default (bug#41250)

[Juri Linkov]

> after this commit I can no longer do a clean build:
>
> % git clean -xfd
> % ./autogen.sh
> % ./configure --prefix=/usr/local/emacs --with-x-toolkit=lucid --with-mailutils
> % make
>
>   GEN      ../lisp/leim/ja-dic/ja-dic.el
>   INFO     Processing OKURI-ARI entries
>   INFO     Processing POSTFIX entries
> Loading macroexp.elc...
> Invalid read syntax: "?"
> make[3]: *** [Makefile:143: ../lisp/leim/ja-dic/ja-dic.el] Error 255
> make[3]: Leaving directory '/home/herbert/source/emacs/leim'
> make[2]: *** [Makefile:352: leim] Error 2
> make[2]: Leaving directory '/home/herbert/source/emacs/lisp'
> make[1]: *** [Makefile:794: ../lisp/loaddefs.el] Error 2
> make[1]: Leaving directory '/home/herbert/source/emacs/src'
> make: *** [Makefile:424: src] Error 2

The problem is that the source file mule-util.el can't use the Unicode
character '…'.  Please see http://debbugs.gnu.org/41250 for the discussion.

This is fixed now.

bug#43822: 28.0.50; build failed on revno.5ec2115

[Juri Linkov]

>>     Use '…' for ellipsis in truncate-string-to-width by default (bug#41250)
>
> Probably because of this change, building Emacs got to fail (on
> at least Cygwin platform).  Before the change (96a8e8) it's ok.
> An output of make is below.  Where I added (setq debug-on-error t)
> to `skkdic-convert-postfix' in "lisp/international/ja-dic-cnv.el".
>
> Note that Emacs tries to load the file named "...".  I couldn't
> find how `lookup-nested-alist' calls `load-with-code-conversion',
> sorry.

This breakage is fixed now.  Please see bug#41250 for the discussion.
This bug could be left open until a more general solution is found.

Re: master 5ec2115: Use '…' for ellipsis in truncate-string-to-width by default (bug#41250)

[Herbert J. Skuhra]

On Tue, Oct 06, 2020 at 10:58:32AM +0300, Juri Linkov wrote:
> > after this commit I can no longer do a clean build:
> >
> > % git clean -xfd
> > % ./autogen.sh
> > % ./configure --prefix=/usr/local/emacs --with-x-toolkit=lucid --with-mailutils
> > % make
> >
> >   GEN      ../lisp/leim/ja-dic/ja-dic.el
> >   INFO     Processing OKURI-ARI entries
> >   INFO     Processing POSTFIX entries
> > Loading macroexp.elc...
> > Invalid read syntax: "?"
> > make[3]: *** [Makefile:143: ../lisp/leim/ja-dic/ja-dic.el] Error 255
> > make[3]: Leaving directory '/home/herbert/source/emacs/leim'
> > make[2]: *** [Makefile:352: leim] Error 2
> > make[2]: Leaving directory '/home/herbert/source/emacs/lisp'
> > make[1]: *** [Makefile:794: ../lisp/loaddefs.el] Error 2
> > make[1]: Leaving directory '/home/herbert/source/emacs/src'
> > make: *** [Makefile:424: src] Error 2
> 
> The problem is that the source file mule-util.el can't use the Unicode
> character '…'.  Please see http://debbugs.gnu.org/41250 for the discussion.
> 
> This is fixed now.

Thanks. But now I get:

Loading macroexp.elc...
Invalid read syntax: "\\N{HORIZONTAL ELLIPSIS}"
make[3]: *** [../lisp/leim/ja-dic/ja-dic.el] Error 255
make[2]: *** [leim] Error 2
make[1]: *** [../lisp/loaddefs.el] Error 2
make: *** [src] Error 2

-- 
Herbert

Re: master 5ec2115: Use '…' for ellipsis in truncate-string-to-width by default (bug#41250)

[Stefan Monnier]

> The problem is that the source file mule-util.el can't use the Unicode
> character '…'.

Yes, it can.  It's some other code which messes things up.


        Stefan

RE: master 5ec2115: Use '…' for ellipsis in truncate-string-to-width by default (bug#41250)

[Drew Adams]

> The problem is that the source file mule-util.el can't use the Unicode
> character '…'.  Please see
> https://urldefense.com/v3/__http://debbugs.gnu.org/41250__;!!GqivPVa7Brio!Myw
> 9NLEyEGq9q15InHIL4DosZcvKPlrhRBDjGAthnl2obqktBvxncP_YLpQWXegw$  for the
> discussion.

FWIW, in that bug (#41250) thread, I expressed 
the opinion about the ellipsis that we should
use "..." and not "…", which is not so readable.

The ellipsis is used to indicate truncation.
By using "..." we include two more chars, but
it suffices to truncate two additional chars,
to end with the same width of text.  (And users
can control the max width by an option.)

https://debbugs.gnu.org/cgi/bugreport.cgi?bug=41250#144

That opinion was ignored, and instead hoops were
jumped through to try to use "…".  IMO, not only
is it not worth trying to do that, the result,
if successful, is worse.

That "…" char is pretty useless, in general.
Its appearance has nothing in common with that
of a real ellipsis, at least when a fixed-width
font is used.

Re: master 5ec2115: Use '…' for ellipsis in truncate-string-to-width by default (bug#41250)

[Stefan Monnier]

> That "…" char is pretty useless, in general.
> Its appearance has nothing in common with that
> of a real ellipsis, at least when a fixed-width
> font is used.

FWIW, I've been using "…" for that ellipsis for a long time now, and
I use a fixed-width font (misc-fixed-semicondensed) pretty
much everywhere.  So obviously my opinion disagrees with yours.

I did end up using "……" at a few places (mostly for the
outline-minor-mode ellipses), tho, because while I find "..." too long,
"…" ended up a bit too short for my taste in that case.

For `truncate-string-to-width` (which is also used to truncate elements
which are significantly shorter than a line's width (contrary to the
outline-minor-mode ellipses) typically in multi-column tabular data),
I found the extra 2 chars very welcome.


        Stefan

RE: master 5ec2115: Use '…' for ellipsis in truncate-string-to-width by default (bug#41250)

[Drew Adams]

> > That "…" char is pretty useless, in general.
> > Its appearance has nothing in common with that
> > of a real ellipsis, at least when a fixed-width
> > font is used.
> 
> FWIW, I've been using "…" for that ellipsis for a long time now, and
> I use a fixed-width font (misc-fixed-semicondensed) pretty
> much everywhere.  So obviously my opinion disagrees with yours.
> 
> I did end up using "……" at a few places (mostly for the
> outline-minor-mode ellipses), tho, because while I find "..." too long,
> "…" ended up a bit too short for my taste in that case.
> 
> For `truncate-string-to-width` (which is also used to truncate elements
> which are significantly shorter than a line's width (contrary to the
> outline-minor-mode ellipses) typically in multi-column tabular data),
> I found the extra 2 chars very welcome.

OK.

There's an argument for users to be able to customize
the string, beyond just the overall length/width.

bug#43822: 28.0.50; build failed on revno.5ec2115

[Juri Linkov]

tags 43822 fixed
close 43822 28.0.50
quit

>>>     Use '…' for ellipsis in truncate-string-to-width by default (bug#41250)
>>
>> Probably because of this change, building Emacs got to fail (on
>> at least Cygwin platform).  Before the change (96a8e8) it's ok.
>> An output of make is below.  Where I added (setq debug-on-error t)
>> to `skkdic-convert-postfix' in "lisp/international/ja-dic-cnv.el".
>>
>> Note that Emacs tries to load the file named "...".  I couldn't
>> find how `lookup-nested-alist' calls `load-with-code-conversion',
>> sorry.
>
> This breakage is fixed now.  Please see bug#41250 for the discussion.
> This bug could be left open until a more general solution is found.

Stefan fixed this in bug#41250, so closing this report as well.

bug#43822: 28.0.50; build failed on revno.5ec2115

[Katsumi Yamaoka]

On Tue, 06 Oct 2020 11:04:22 +0300, Juri Linkov wrote:
> This breakage is fixed now.  Please see bug#41250 for the discussion.
> This bug could be left open until a more general solution is found.

Oh, I understood it to be due to misplaced let-bound
`coding-system-for-read' that prevents mule-util.el, in which
"…" is added recently, from being autoloaded.

On Tue, 06 Oct 2020 21:27:26 +0300, Juri Linkov wrote:
> Stefan fixed this in bug#41250, so closing this report as well.

Works fine now.  Thanks.