NEWS

Branch :
Show log
Commit
Author : Bruno Haible
Date : 2024-12-15 10:37:13
Hash : ef2f457f
Message : Support multiple suffixes in iconv_open, like glibc does. * lib/iconv_open1.h: Use a loop when looking for the common suffixes. * tests/test-discard.c: New file. * tests/Makefile.in (check): Run test-discard. (test-discard, test-discard.@OBJEXT@): New targets. (clean): Remove test-discard. (SOURCE_FILES): Add test-discard.c. * NEWS: Mention the change.
NEWS
New in 1.18:
* Many more transliterations, in particular also of Emoji characters.
* The iconv_open function and the iconv program now support multiple suffixes,
  such as //TRANSLIT//IGNORE, not only one.
* GB18030 is now an alias for GB18030:2005. A new converter for GB18030:2022
  is added. Since this encoding merely cleans up a few private-use-area
  mappings, you can continue to use the GB18030 converter, for backward
  compatibility. Its Unicode to GB18030 conversion direction has been
  enhanced, to help transitioning away from PUA code points.
* When converting from/to an EBCDIC encoding, a non-standard way of
  converting newlines can be requested
    - at the C level, by calling iconvctl with argument ICONV_SET_FROM_SURFACE
      or ICONV_SET_TO_SURFACE, or
    - from the iconv program, by setting the environment variable
      ICONV_EBCDIC_ZOS_UNIX to a non-empty value.
* Special support for z/OS: The iconv program adds a charset metadata tag to
  its output file. (Contributed by Mike Fulton.)
* For conversions from UCS-2, UCS-4, UTF-16, UTF-32, invoking
  iconv(cd,NULL,NULL,...) now preserves the byte order state.

New in 1.17:
* The libiconv library is now licensed under the LGPL version 2.1, instead of
  the LGPL version 2.0. The iconv program continues to be licensed under GPL
  version 3.
* Added converters for many single-byte EBCDIC encodings:
  IBM-{037,273,277,278,280,282,284,285,297,423,424,425,500,838,870,871,875},
  IBM-{880,905,924,1025,1026,1047,1097,1112,1122,1123,1130,1132,1137,1140},
  IBM-{1141,1142,1143,1144,1145,1146,1147,1148,1149,1153,1154,1155,1156,1157},
  IBM-{1158,1160,1164,1165,1166,4971,12712,16804}.
  They are available through the configure option '--enable-extra-encodings'.

New in 1.16:
* The preloadable library has been removed.

New in 1.15:
* The UTF-8 converter now rejects surrogates and out-of-range code points.
* Added ISO-2022-JP-MS converter.
* Updated the CP1255 converter to map one more character.
* The functions now support strings longer than 2 GB.

New in 1.14:
* The 'iconv' program now produces its output as soon as it can. It no longer
  unnecessarily waits for more input.
* Updated the GB18030 converter to map 25 characters to code points that have
  been to Unicode since 2000, rather than to code points in the Private Use
  Area.
* Updated the BIG5-HKSCS converter. The old BIG5-HKSCS converter is renamed to
  BIG5-HKSCS:2004. A new converter BIG5-HKSCS:2008 is added. BIG5-HKSCS is now
  an alias for BIG5-HKSCS:2008.
* Fixed a bug in the conversion to wchar_t.
* Fixed a small bug in the CP1258 converter.

New in 1.13:
* The library and the iconv program now understand platform dependent aliases,
  for better compatibility with the platform's own iconv_open function.
  Examples: "646" on Solaris, "iso88591" on HP-UX, "IBM-1252" on AIX.
* For stateful encodings, when the input ends with a shift sequence followed
  by invalid input, the iconv function now increments the input pointer past
  the shift sequence before returning (size_t)(-1) with errno = EILSEQ. This
  is also like GNU libc's iconv() behaves.
* The library exports a new function iconv_open_into() that stores the
  conversion descriptor in pre-allocated memory, rather than allocating fresh
  memory for it.
* Added CP1131 converter.

New in 1.12:
* The iconv program is now licensed under the GPL version 3, instead of the
  GPL version 2. The libiconv library continues to be licensed under LGPL.
* Added RK1048 converter.
* On AIX, an existing system libiconv no longer causes setlocale() to fail.
* Upgraded EUC-KR, JOHAB to include the Korean postal code sign.

New in 1.11:
* The iconv program has new options --unicode-subst, --byte-subst,
  --widechar-subst that allow to specify substitutions for characters that
  cannot be converted.
* The iconv program now understands long options:
    long option    equivalent to
    --from-code    -f
    --to-code      -t
    --list         -l
    --silent       -s
* The CP936 converter is now different from the GBK converter: it has changed
  to include the Euro sign and private area characters. CP936 is no longer an
  alias of GBK.
* Updated GB18030 converter to include all private area characters.
* Updated CP950 converter to include the Euro sign and private area characters.
* Updated CP949 converter to include private area characters.
* Updated the BIG5-HKSCS converter. The old BIG5-HKSCS converter is renamed to
  BIG5-HKSCS:1999 and updated to Unicode 4. New converters BIG5-HKSCS:2001 and
  BIG5-HKSCS:2004 are added. BIG5-HKSCS is now an alias for BIG5-HKSCS:2004.
* Added a few irreversible mappings to the CP932 converter.
* Tidy up the list of symbols exported from libiconv (assumes gcc >= 4.0).

New in 1.10:
* Added ISO-8859-11 converter.
* Updated the ISO-8859-7 converter.
* Added ATARIST converter, available through --enable-extra-encodings.
* Added BIG5-2003 converter (experimental), available through
  --enable-extra-encodings.
* Updated EUC-TW converter to include the Euro sign.
* The preloadable library has been renamed from libiconv_plug.so to
  preloadable_libiconv.so.
* Portability to mingw.

New in 1.9:
* Many more transliterations.
* New configuration option --enable-relocatable.  See the INSTALL.generic file
  for details.

New in 1.8:
* The iconv program has new options -l, -c, -s.
* The iconv program is internationalized.
* Added C99 converter.
* Added KOI8-T converter.
* New configuration option --enable-extra-encodings that enables a bunch of
  additional encodings; see the README for details.
* Updated the ISO-8859-16 converter.
* Upgraded BIG5-HKSCS, EUC-TW, ISO-2022-CN, ISO-2022-CN-EXT converters to
  Unicode 3.2.
* Upgraded EUC-KR, CP949, JOHAB converters to include the Euro sign.
* Changed the ARMSCII-8 converter.
* Extended the EUC-JP encoder so that YEN SIGN characters don't cause failures
  in Shift_JIS to EUC-JP conversion.
* The JAVA converter now handles characters outside the Unicode BMP correctly.
* Fixed a bug in the CP1255, CP1258, TCVN decoders: The base characters of
  combining characters could be dropped at the end of the conversion buffer.
* Fixed a bug in the transliteration that could lead to excessive memory
  allocations in libintl when transliteration was needed.
* Portability to BSD/OS and SCO 3.2.5.

New in 1.7:
* Added UTF-32, UTF-32BE, UTF-32LE converters.
* Changed CP1255, CP1258 and TCVN converters to handle combining characters.
* Changed EUC-JP, SHIFT_JIS, CP932, ISO-2022-JP, ISO-2022-JP-2, ISO-2022-JP-1
  converters to use fullwidth Yen sign instead of halfwidth Yen sign, and
  fullwidth tilde instead of halfwidth tilde.
* Upgraded EUC-TW, ISO-2022-CN, ISO-2022-CN-EXT converters to Unicode 3.1.
* Changed the GB18030 converter to not reject unassigned and private-use
  Unicode characters.
* Fixed a bug in the byte order mark treatment of the UCS-4 decoder.
* The manual pages are now distributed also in HTML format.

New in 1.6:
* The iconv program's -f and -t options are now optional.
* Many more transliterations.
* Added CP862 converter.
* Changed the GB18030 converter.
* Portability to DOS with DJGPP.

New in 1.5:
* Added an iconv(1) program.
* New locale dependent encodings "char", "wchar_t".
* Transliteration is now off by default. Use a //TRANSLIT suffix to enable it.
* The JOHAB encoding is documented again.
* Changed a few mappings in the CP950 converter.

New in 1.4:
* Added GB18030, BIG5HKSCS converters.
* Portability to OS/2 with emx+gcc.

New in 1.3:
* Added UCS-2BE, UCS-2LE, UCS-4BE, UCS-4LE converters.
* Fixed the definition of EILSEQ on SunOS4.
* Fixed a build problem on OSF/1.
* Support for building as a shared library on Woe32.

New in 1.2:
* Added UTF-16BE and UTF-16LE converters.
* Changed the UTF-16 encoder.
* Fixed the treatment of tab characters in the UTF-7 converter.
* Fixed an internal error when output buffer was not large enough.

New in 1.1:
* Added ISO-8859-16 converter.
* Added CP932 converter, a variant of SHIFT_JIS.
* Added CP949 converter, a variant of EUC-KR.
* Improved the ISO-2022-CN-EXT converter: It now covers the ISO-IR-165 range.
* Updated the ISO-8859-8 conversion table.
* The JOHAB encoding is deprecated and not documented any more.
* Fixed two build problems: 1. "make -n check" failed. 2. When libiconv was
  already installed, "make" failed.

New in 1.0:
* Added transliteration facilities.
* Added a test suite.
* Fixed the iconv(3) manual page and function: the return value was not
  described correctly.
* Fixed a bug in the CP1258 decoder: invalid bytes now yield EILSEQ instead of
  U+FFFD.
* Fixed a bug in the Georgian-PS encoder: accept U+00E6.
* Fixed a bug in the EUC-JP encoder: reject 0x8E5C and 0x8E7E.
* Fixed a bug in the KSC5601 and JOHAB converters: they recognized some Hangul
  characters at some invalid code positions.
* Fixed a bug in the EUC-TW decoder; it was severely broken.
* Fixed a bug in the CP950 converter: it recognized a dubious BIG5 range.

New in 0.3:
* Reduced the size of the tables needed for the JOHAB converter.
* Portability to Woe32.

New in 0.2:
* Added KOI8-RU, CP850, CP866, CP874, CP950, ISO-2022-CN-EXT, GBK and
  ISO-2022-JP-1 converters.
* Added MACINTOSH as an alias for MAC-ROMAN.
* Added ASMO-708 as an alias for ISO-8859-6.
* Added ELOT_928 as an alias for ISO-8859-7.
* Improved the EUC-TW converter: Treat CNS 11643 plane 3.
* Improved the ISO-2022-KR and EUC-KR converters: Hangul characters are
  decomposed into Jamo when needed.
* Improved the CP932 converter.
* Updated the CP1133, MULELAO-1 and ARMSCII-8 mappings.
* The EUC-JP and SHIFT_JIS converters now cover the user-defined range.
* Fixed a possible buffer overrun in the JOHAB converter.
* Fixed a bug in the UTF-7, ISO-2022-*, HZ decoders: a shift sequence a the
  end of the input no longer gives an error.
* The HZ encoder now always terminates its output in the ASCII state.
* Use a perfect hash table for looking up the aliases.

New in 0.1:
* Portability to Linux/glibc-2.0.x, Linux/libc5, OSF/1, FreeBSD.
* Fixed a bug in the EUC-JP decoder. Extended the ISO-2022-JP-2 converter.
* Made TIS-620 mapping consistent with glibc-2.1.
thodg/libiconv/NEWS

Commit

thodg/libiconv /NEWS
