code.delx.au - gnu-emacs/blob - admin/notes/unicode

   1                                             -*-mode: text; coding: utf-8;-*-
   2
   3 Copyright (C) 2002-2016 Free Software Foundation, Inc.
   4 See the end of the file for license conditions.
   5
   6 Importing a new Unicode Standard version into Emacs
   7 -------------------------------------------------------------
   8
   9 Emacs uses the following files from the Unicode Character Database
  10 (a.k.a. "UCD):
  11
  12   . UnicodeData.txt
  13   . Blocks.txt
  14   . BidiMirroring.txt
  15   . BidiBrackets.txt
  16   . IVD_Sequences.txt
  17   . NormalizationTest.txt
  18
  19 First, these files need to be copied into admin/unidata/, and then
  20 Emacs should be rebuilt for them to take effect.  Rebuilding Emacs
  21 updates several derived files elsewhere in the Emacs source tree,
  22 mainly in lisp/international/.
  23
  24 When Emacs is rebuilt for the first time after importing the new
  25 files, pay attention to any warning or error messages.  In particular,
  26 admin/unidata/unidata-gen.el will complain if UnicodeData.txt defines
  27 new bidirectional attributes of characters, because unidata-gen.el,
  28 bidi.c and dispextern.h need to be updated in that case; failure to do
  29 so will cause aborts in redisplay.
  30
  31 Next, review the changes in UnicodeData.txt vs the previous version
  32 used by Emacs.  Any changes, be it introduction of new scripts or
  33 addition of codepoints to existing scripts, might need corresponding
  34 changes in the data used for filling the category-table, case-table,
  35 and char-width-table.  The additional scripts should cause automatic
  36 updates in charscript.el, but it is a good idea to look at the results
  37 and see if any changes in admin/unidata/blocks.awk are required.
  38
  39 Any new scripts added by UnicodeData.txt will also need updates to
  40 script-representative-chars defined in fontset.el, and also the list
  41 of OTF script tags in otf-script-alist, whose source is on this page:
  42
  43   https://www.microsoft.com/typography/otspec/scripttags.htm
  44
  45 Other databases in fontset.el might also need to be updated as needed.
  46
  47 The function 'ucs-names', defined in lisp/international/mule-cmds.el,
  48 might need to be updated because it knows about used and unused ranges
  49 of Unicode codepoints, which a new release of the Unicode Standard
  50 could change.
  51
  52 Finally, test normalization functions against NormalizationTests.txt,
  53 in the test/ directory run:
  54
  55   make lisp/international/ucs-normalize-tests
  56
  57 See commentary in test/lisp/international/ucs-normalize-tests.el
  58 regarding failing lines.
  59
  60 Problems, fixmes and other unicode-related issues
  61 -------------------------------------------------------------
  62
  63 Notes by fx to record various things of variable importance.  Handa
  64 needs to check them -- don't take too seriously, especially with
  65 regard to completeness.
  66
  67  * SINGLE_BYTE_CHAR_P returns true for Latin-1 characters, which has
  68    undesirable effects.  E.g.:
  69    (multibyte-string-p (let ((s "x")) (aset s 0 ?£) s)) => nil
  70    (multibyte-string-p (concat [?£])) => nil
  71    (text-char-description ?£) => "M-#"
  72
  73         These examples are all fixed by the change of 2002-10-14, but
  74         there still exist questionable SINGLE_BYTE_CHAR_P in the
  75         code (keymap.c and print.c).
  76
  77  * Rationalize character syntax and its relationship to the Unicode
  78    database.  (Applies mainly to symbol an punctuation syntax.)
  79
  80  * Fontset handling and customization needs work.  We want to relate
  81    fonts to scripts, probably based on the Unicode blocks.  The
  82    presence of small-repertoire 10646-encoded fonts in XFree 4 is a
  83    pain, not currently worked round.
  84
  85         With the change on 2002-07-26, multiple fonts can be
  86         specified in a fontset for a specific range of characters.
  87         Each range can also be specified by script.  Before using
  88         ISO10646 fonts, Emacs checks their repertories to avoid such
  89         fonts that don't have a glyph for a specific character.
  90
  91         fx has worked on fontset customization, but was stymied by
  92         basic problems with the way the default face is dealt with
  93         (and something else, I think).  This needs revisiting.
  94
  95  * Work is also needed on charset and coding system priorities.
  96
  97  * The relevant bits of latin1-disp.el need porting (and probably
  98    re-naming/updating).  See also cyril-util.el.
  99
 100  * Quail files need more work now the encoding is largely irrelevant.
 101
 102  * What to do with the old coding categories stuff?
 103
 104  * The preferred-coding-system property of charsets should probably be
 105    junked unless it can be made more useful now.
 106
 107  * find-multibyte-characters needs looking at.
 108
 109  * Implement Korean cp949/UHC, BIG5-HKSCS and any other important missing
 110    charsets.
 111
 112  * Lazy-load tables for unify-charset somehow?
 113
 114         Actually, Emacs clears out all charset maps and unify-map just
 115         before dumping, and they are loaded again on demand by the
 116         dumped emacs.  But, those maps (char tables) generated while
 117         temacs is running can't be removed from the dumped emacs.
 118
 119  * iso-2022 charsets get unified on i/o.
 120
 121         With the change on 2003-01-06, decoding routines put the 'charset'
 122         property onto decoded text, and iso-2022 encoder pay attention
 123         to it.  Thus, for instance, reading and writing by
 124         iso-2022-7bit preserve the original designation sequences.
 125         The property name 'preferred-charset' may be better?
 126
 127         We may have to utilize this property to decide a font.
 128
 129  * Revisit locale processing: look at treating the language and
 130    charset parts separately.  (Language should affect things like
 131    spelling and calendar, but that's not a Unicode issue.)
 132
 133  * Handle Unicode combining characters usefully, e.g. diacritics, and
 134    handle more scripts specifically (à la Devanagari).  There are
 135    issues with canonicalization.
 136
 137  * We need tabular input methods, e.g. for maths symbols.  (Not
 138    specific to Unicode.)
 139
 140  * Need multibyte text in menus, e.g. for the above.  (Not specific to
 141    Unicode -- see Emacs etc/TODO, but now mostly works with gtk.)
 142
 143  * Populate char-width-table correctly for Unicode characters and
 144    worry about what happens when double-width charsets covering
 145    non-CJK characters are unified.
 146
 147  * There are type errors lurking, e.g. in
 148    Fcheck_coding_systems_region.  Define ENABLE_CHECKING to find them.
 149
 150  * Old auto-save files, and similar files, such as Gnus drafts,
 151    containing non-ASCII characters probably won't be re-read correctly.
 152
 153
 154 Source file encoding
 155 --------------------
 156
 157 Most Emacs source files are encoded in UTF-8 (or in ASCII, which is a
 158 subset), but there are a few exceptions, listed below.  Perhaps
 159 someday many of these files will be converted to UTF-8, for
 160 convenience when using tools like 'grep -r', but this might need
 161 nontrivial changes to the build process.
 162
 163  * chinese-big5
 164
 165      These are verbatim copies of files taken from external sources.
 166      They haven't been converted to UTF-8.
 167
 168         leim/CXTERM-DIC/4Corner.tit
 169         leim/CXTERM-DIC/ARRAY30.tit
 170         leim/CXTERM-DIC/ECDICT.tit
 171         leim/CXTERM-DIC/ETZY.tit
 172         leim/CXTERM-DIC/PY-b5.tit
 173         leim/CXTERM-DIC/Punct-b5.tit
 174         leim/CXTERM-DIC/QJ-b5.tit
 175         leim/CXTERM-DIC/ZOZY.tit
 176         leim/MISC-DIC/CTLau-b5.html
 177         leim/MISC-DIC/cangjie-table.b5
 178
 179  * chinese-iso-8bit
 180
 181      These are verbatim copies of files taken from external sources.
 182      They haven't been converted to UTF-8.
 183
 184         leim/CXTERM-DIC/CCDOSPY.tit
 185         leim/CXTERM-DIC/Punct.tit
 186         leim/CXTERM-DIC/QJ.tit
 187         leim/CXTERM-DIC/SW.tit
 188         leim/CXTERM-DIC/TONEPY.tit
 189         leim/MISC-DIC/CTLau.html
 190         leim/MISC-DIC/pinyin.map
 191         leim/MISC-DIC/ziranma.cin
 192
 193  * cp850
 194
 195      This file contains non-ASCII characters in unibyte strings.  When
 196      editing a keyboard layout it's more convenient to see 'é' than
 197      '\202', and the MS-DOS compiler requires the single byte if a
 198      backslash escape is not being used.
 199
 200         src/msdos.c
 201
 202  * iso-2022-cn-ext
 203
 204      This file is externally generated from leim/MISC-DIC/cangjie-table.b5
 205      by Big5->CNS converter.  It hasn't been converted to UTF-8.
 206
 207         leim/MISC-DIC/cangjie-table.cns
 208
 209  * japanese-iso-8bit
 210
 211      SKK-JISYO.L is a verbatim copy of a file taken from an external source.
 212      It hasn't been converted to UTF-8.
 213
 214         leim/SKK-DIC/SKK-JISYO.L
 215
 216  * japanese-shift-jis
 217
 218      This is a verbatim copy of a file taken from an external source.
 219      It hasn't been converted to UTF-8.
 220
 221         admin/charsets/mapfiles/cns2ucsdkw.txt
 222
 223  * iso-2022-7bit
 224
 225      This file switches between CJK charsets, which is not encoded in UTF-8.
 226
 227         etc/HELLO
 228
 229      Each of these files contains just one CJK charset, but Emacs
 230      currently has no easy way to specify set-charset-priority on a
 231      per-file basis, so converting any of these files to UTF-8 might
 232      change the file's appearance when viewed by an Emacs that is
 233      operating in some other language environment.
 234
 235         etc/tutorials/TUTORIAL.ja
 236         lisp/international/ja-dic-cnv.el
 237         lisp/international/ja-dic-utl.el
 238         lisp/international/kinsoku.el
 239         lisp/international/kkc.el
 240         lisp/international/titdic-cnv.el
 241         lisp/language/japan-util.el
 242         lisp/language/japanese.el
 243         lisp/leim/quail/cyril-jis.el
 244         lisp/leim/quail/hanja-jis.el
 245         lisp/leim/quail/japanese.el
 246         lisp/leim/quail/py-punct.el
 247         lisp/leim/quail/pypunct-b5.el
 248
 249      This file contains just Chinese characters, and has same problem.
 250      Also, it contains characters that cannot be encoded in UTF-8.
 251
 252         lisp/international/titdic-cnv.el
 253
 254  * utf-8-emacs
 255
 256      These files contain characters that cannot be encoded in UTF-8.
 257
 258         lisp/language/ethio-util.el
 259         lisp/language/ethiopic.el
 260         lisp/language/ind-util.el
 261         lisp/language/tibet-util.el
 262         lisp/language/tibetan.el
 263         lisp/leim/quail/ethiopic.el
 264         lisp/leim/quail/tibetan.el
 265
 266  * binary files
 267
 268      These files contain binary data, and are not text files.
 269      Some of the entries in this list are patterns, and stand for any
 270      files with the listed extension.
 271
 272         *.gz
 273         *.icns
 274         *.ico
 275         *.pbm
 276         *.pdf
 277         *.png
 278         *.sig
 279         etc/e/eterm-color
 280         etc/package-keyring.gpg
 281         msdos/emacs.pif
 282         nextstep/GNUstep/Emacs.base/Resources/emacs.tiff
 283         nt/icons/hand.cur
 284
 285 \f
 286 This file is part of GNU Emacs.
 287
 288 GNU Emacs is free software: you can redistribute it and/or modify
 289 it under the terms of the GNU General Public License as published by
 290 the Free Software Foundation, either version 3 of the License, or
 291 (at your option) any later version.
 292
 293 GNU Emacs is distributed in the hope that it will be useful,
 294 but WITHOUT ANY WARRANTY; without even the implied warranty of
 295 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 296 GNU General Public License for more details.
 297
 298 You should have received a copy of the GNU General Public License
 299 along with GNU Emacs.  If not, see <http://www.gnu.org/licenses/>.