emacs/admin/notes/unicode

                                            -*-mode: text; coding: utf-8;-*-

Copyright (C) 2002-2021 Free Software Foundation, Inc.
See the end of the file for license conditions.

Importing a new Unicode Standard version into Emacs
-------------------------------------------------------------

Emacs uses the following files from the Unicode Character Database
(a.k.a. "UCD):

  . UnicodeData.txt
  . Blocks.txt
  . BidiBrackets.txt
  . BidiCharacterTest.txt
  . BidiMirroring.txt
  . IVD_Sequences.txt
  . NormalizationTest.txt
  . SpecialCasing.txt

First, the first 7 files need to be copied into admin/unidata/, and
the file https://www.unicode.org/copyright.html should be copied over
copyright.html in admin/unidata (that file might need trailing
whitespace removed before it can be committed to the Emacs
repository).

Then Emacs should be rebuilt for them to take effect.  Rebuilding
Emacs updates several derived files elsewhere in the Emacs source
tree, mainly in lisp/international/.

When Emacs is rebuilt for the first time after importing the new
files, pay attention to any warning or error messages.  In particular,
admin/unidata/unidata-gen.el will complain if UnicodeData.txt defines
new bidirectional attributes of characters, because unidata-gen.el,
bidi.c and dispextern.h need to be updated in that case; failure to do
so will cause aborts in redisplay.  unidata-gen.el will also complain
if the format of the Unicode Copyright notice in copyright.html
changed in significant ways; in that case, update the regular
expression in unidata-gen-file used to extract the copyright string.

Next, review the changes in UnicodeData.txt vs the previous version
used by Emacs.  Any changes, be it introduction of new scripts or
addition of codepoints to existing scripts, might need corresponding
changes in the data used for filling the category-table, case-table,
and char-width-table.  The additional scripts should cause automatic
updates in charscript.el, but it is a good idea to look at the results
and see if any changes in admin/unidata/blocks.awk are required.

The setting of char-width-table around line 1200 of characters.el
should be checked against the latest version of the Unicode file
EastAsianWidth.txt, and any discrepancies fixed: double-width
characters are those marked with W or F in that file.  Zero-width
characters are not taken from EastAsianWidth.txt, they are those whose
Unicode General Category property is one of Mn, Me, or Cf, and also
Hangul jungseong and jongseong characters (a.k.a. "Jamo medial vowels"
and "Jamo final consonants").

Any new scripts added by UnicodeData.txt will also need updates to
script-representative-chars defined in fontset.el, and also the list
of OTF script tags in otf-script-alist, whose source is on this page:

  https://docs.microsoft.com/en-us/typography/opentype/spec/scripttags

Other databases in fontset.el might also need to be updated as needed.
One notable place to check is the function setup-default-fontset,
where new scripts will generally need some addition, most probably to
the list of "simple" scripts (search for "Simple").

The function 'ucs-names', defined in lisp/international/mule-cmds.el,
might need to be updated because it knows about used and unused ranges
of Unicode codepoints, which a new release of the Unicode Standard
could change.

Next, test normalization functions against NormalizationTests.txt,
in the test/ directory run:

  make lisp/international/ucs-normalize-tests

See commentary in test/lisp/international/ucs-normalize-tests.el
regarding failing lines.

The file BidiCharacterTest.txt should be copied to the test suite, and
if its format has changed, the file biditest.el there should be
modified to follow suit.

Problems, fixmes and other unicode-related issues
-------------------------------------------------------------

Notes by fx to record various things of variable importance.  Handa
needs to check them -- don't take too seriously, especially with
regard to completeness.

 * SINGLE_BYTE_CHAR_P returns true for Latin-1 characters, which has
   undesirable effects.  E.g.:
   (multibyte-string-p (let ((s "x")) (aset s 0 ?£) s)) => nil
   (multibyte-string-p (concat [?£])) => nil
   (text-char-description ?£) => "M-#"

	These examples are all fixed by the change of 2002-10-14, but
	there still exist questionable SINGLE_BYTE_CHAR_P in the
	code (keymap.c and print.c).

 * Rationalize character syntax and its relationship to the Unicode
   database.  (Applies mainly to symbol and punctuation syntax.)

 * Fontset handling and customization needs work.  We want to relate
   fonts to scripts, probably based on the Unicode blocks.  The
   presence of small-repertoire 10646-encoded fonts in XFree 4 is a
   pain, not currently worked round.

	With the change on 2002-07-26, multiple fonts can be
	specified in a fontset for a specific range of characters.
	Each range can also be specified by script.  Before using
	ISO10646 fonts, Emacs checks their repertories to avoid such
	fonts that don't have a glyph for a specific character.

	fx has worked on fontset customization, but was stymied by
	basic problems with the way the default face is dealt with
	(and something else, I think).  This needs revisiting.

 * Work is also needed on charset and coding system priorities.

 * The relevant bits of latin1-disp.el need porting (and probably
   re-naming/updating).  See also cyril-util.el.

 * Quail files need more work now the encoding is largely irrelevant.

 * What to do with the old coding categories stuff?

 * The preferred-coding-system property of charsets should probably be
   junked unless it can be made more useful now.

 * find-multibyte-characters needs looking at.

 * Implement Korean cp949/UHC, BIG5-HKSCS and any other important missing
   charsets.

 * Lazy-load tables for unify-charset somehow?

	Actually, Emacs clears out all charset maps and unify-map just
	before dumping, and they are loaded again on demand by the
	dumped emacs.  But, those maps (char tables) generated while
	temacs is running can't be removed from the dumped emacs.

 * iso-2022 charsets get unified on i/o.

	With the change on 2003-01-06, decoding routines put the 'charset'
	property onto decoded text, and iso-2022 encoder pay attention
	to it.  Thus, for instance, reading and writing by
	iso-2022-7bit preserve the original designation sequences.
	The property name 'preferred-charset' may be better?

	We may have to utilize this property to decide a font.

 * Revisit locale processing: look at treating the language and
   charset parts separately.  (Language should affect things like
   spelling and calendar, but that's not a Unicode issue.)

 * Handle Unicode combining characters usefully, e.g. diacritics, and
   handle more scripts specifically (à la Devanagari).  There are
   issues with canonicalization.

 * We need tabular input methods, e.g. for maths symbols.  (Not
   specific to Unicode.)

 * Need multibyte text in menus, e.g. for the above.  (Not specific to
   Unicode -- see Emacs etc/TODO, but now mostly works with gtk.)

 * Populate char-width-table correctly for Unicode characters and
   worry about what happens when double-width charsets covering
   non-CJK characters are unified.

 * There are type errors lurking, e.g. in
   Fcheck_coding_systems_region.  Define ENABLE_CHECKING to find them.

 * Old auto-save files, and similar files, such as Gnus drafts,
   containing non-ASCII characters probably won't be re-read correctly.


Source file encoding
--------------------

Most Emacs source files are encoded in UTF-8 (or in ASCII, which is a
subset), but there are a few exceptions, listed below.  Perhaps
someday many of these files will be converted to UTF-8, for
convenience when using tools like 'grep -r', but this might need
nontrivial changes to the build process.

 * chinese-big5

     These are verbatim copies of files taken from external sources.
     They haven't been converted to UTF-8.

	leim/CXTERM-DIC/4Corner.tit
	leim/CXTERM-DIC/ARRAY30.tit
	leim/CXTERM-DIC/ECDICT.tit
	leim/CXTERM-DIC/ETZY.tit
	leim/CXTERM-DIC/PY-b5.tit
	leim/CXTERM-DIC/Punct-b5.tit
	leim/CXTERM-DIC/QJ-b5.tit
	leim/CXTERM-DIC/ZOZY.tit
	leim/MISC-DIC/CTLau-b5.html
	leim/MISC-DIC/cangjie-table.b5

 * chinese-iso-8bit

     These are verbatim copies of files taken from external sources.
     They haven't been converted to UTF-8.

	leim/CXTERM-DIC/CCDOSPY.tit
	leim/CXTERM-DIC/Punct.tit
	leim/CXTERM-DIC/QJ.tit
	leim/CXTERM-DIC/SW.tit
	leim/CXTERM-DIC/TONEPY.tit
	leim/MISC-DIC/CTLau.html
	leim/MISC-DIC/pinyin.map
	leim/MISC-DIC/ziranma.cin

 * cp850

     This file contains non-ASCII characters in unibyte strings.  When
     editing a keyboard layout it's more convenient to see 'é' than
     '\202', and the MS-DOS compiler requires the single byte if a
     backslash escape is not being used.

	src/msdos.c

 * iso-2022-cn-ext

     This file is externally generated from leim/MISC-DIC/cangjie-table.b5
     by Big5->CNS converter.  It hasn't been converted to UTF-8.

	leim/MISC-DIC/cangjie-table.cns

 * japanese-iso-8bit

     SKK-JISYO.L is a verbatim copy of a file taken from an external source.
     It hasn't been converted to UTF-8.

	leim/SKK-DIC/SKK-JISYO.L

 * japanese-shift-jis

     This is a verbatim copy of a file taken from an external source.
     It hasn't been converted to UTF-8.

	admin/charsets/mapfiles/cns2ucsdkw.txt

 * iso-2022-jp

     This contains just one CJK charset, but Emacs currently has no
     easy way to specify set-charset-priority on a per-file basis, so
     converting this file to UTF-8 might change the file's appearance
     when viewed by an Emacs that is operating in some other language
     environment.

	etc/tutorials/TUTORIAL.ja

 * iso-2022-7bit

     This file contains multiple Chinese charsets, and converting it
     to UTF-8 would lose the charset property and would change the
     code's behavior.  Although this could be worked around by
     propertizing the strings, that hasn't been done.

	lisp/international/titdic-cnv.el

 * utf-8-emacs

     These files contain characters that cannot be encoded in UTF-8.

	lisp/language/ethio-util.el
	lisp/language/ethiopic.el
	lisp/language/ind-util.el
	lisp/language/tibet-util.el
	lisp/language/tibetan.el
	lisp/leim/quail/ethiopic.el
	lisp/leim/quail/tibetan.el

 * binary files

     These files contain binary data, and are not text files.
     Some of the entries in this list are patterns, and stand for any
     files with the listed extension.

	*.gz
	*.icns
	*.ico
	*.pbm
	*.pdf
	*.png
	*.sig
	etc/e/eterm-color
	etc/package-keyring.gpg
	msdos/emacs.pif
	nextstep/GNUstep/Emacs.base/Resources/emacs.tiff
	nt/icons/hand.cur


This file is part of GNU Emacs.

GNU Emacs is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.

GNU Emacs is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with GNU Emacs.  If not, see <https://www.gnu.org/licenses/>.