Did something Change With Unicode Name Mappings?

Did something Change With Unicode Name Mappings?

[raman]

Symptoms:

C-x  8 RET 
no longer appears  to find things like smiley faces on completion.


Something may have broken in the process of semi-obsoleting
describe-char-unicode-data.

I have variable  describe-char-unicodedata-file set to 
"/usr/local/share/unicode/UnicodeData.txt" --- that file was
downloaded from the Unicode site.

Until recently, names of chars as defined in that file used to be
accepted by C-x 8 RET --- appears to not happen any longer.

--raman 



-- 

-- 

Re: Did something Change With Unicode Name Mappings?

[Eli Zaretskii]

> From: raman@google.com
> Date: Sun, 26 Jun 2016 15:15:34 -0700
> 
> Symptoms:
> 
> C-x  8 RET 
> no longer appears  to find things like smiley faces on completion.

Please give a complete recipe, including the name of the character, or
its part, that you typed.  I tried to reproduce this, but couldn't.

Also, in which version of Emacs do you see that?

> Something may have broken in the process of semi-obsoleting
> describe-char-unicode-data.

AFAIK, "C-x 8 RET" doesn't need any file, since all the data is loaded
into Emacs from the uni-*.el files, which are produced from
UnicodeData.txt at build time.

> I have variable  describe-char-unicodedata-file set to 
> "/usr/local/share/unicode/UnicodeData.txt" --- that file was
> downloaded from the Unicode site.
> 
> Until recently, names of chars as defined in that file used to be
> accepted by C-x 8 RET --- appears to not happen any longer.

It seem to work for me.

Re: Did something Change With Unicode Name Mappings?

[raman]

Eli Zaretskii <eliz@gnu.org> writes:


This was in emacs 25@Head.

The problem went after I did the following:

1. Did a make clean in admin/unicode 
2. Rebuilt Emacs
3. Updated  describe-char-unicodedata-file to point  at the UniCode.txt
in emacs/admin/unicode 


Characters that demonstrated the problem before I did this and that went
away after:

C-x 8 RET  dog tab f tab == "dog face"

🐶

--raman >> From: raman@google.com
>> Date: Sun, 26 Jun 2016 15:15:34 -0700
>> 
>> Symptoms:
>> 
>> C-x  8 RET 
>> no longer appears  to find things like smiley faces on completion.
>
> Please give a complete recipe, including the name of the character, or
> its part, that you typed.  I tried to reproduce this, but couldn't.
>
> Also, in which version of Emacs do you see that?
>
>> Something may have broken in the process of semi-obsoleting
>> describe-char-unicode-data.
>
> AFAIK, "C-x 8 RET" doesn't need any file, since all the data is loaded
> into Emacs from the uni-*.el files, which are produced from
> UnicodeData.txt at build time.
>
>> I have variable  describe-char-unicodedata-file set to 
>> "/usr/local/share/unicode/UnicodeData.txt" --- that file was
>> downloaded from the Unicode site.
>> 
>> Until recently, names of chars as defined in that file used to be
>> accepted by C-x 8 RET --- appears to not happen any longer.
>
> It seem to work for me.

-- 

Re: Did something Change With Unicode Name Mappings?

[Eli Zaretskii]

> From: raman <raman@google.com>
> Cc: emacs-devel@gnu.org
> Date: Mon, 27 Jun 2016 08:04:00 -0700
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> 
> This was in emacs 25@Head.

We recently imported the Unicode data files from the latest version
9.0.0 of the standard, perhaps something went wrong with the build
following that?

> The problem went after I did the following:
> 
> 1. Did a make clean in admin/unicode 
> 2. Rebuilt Emacs
> 3. Updated  describe-char-unicodedata-file to point  at the UniCode.txt
> in emacs/admin/unicode 
> 
> 
> Characters that demonstrated the problem before I did this and that went
> away after:
> 
> C-x 8 RET  dog tab f tab == "dog face"
> 
> 🐶

The completion on character names doesn't use
describe-char-unicodedata-file at all, AFAIK, only the internal
database loaded from uni-*.el files, so I'm guessing that steps 1 and
2 were those which fixed the problem.

Btw, there's no need to point describe-char-unicodedata-file at any
file, Emacs has all that data in its own databases.  And if you are
tracking the master branch, the Unicode database we have there should
always be the latest one anyway.