/ 
Jacode4e-2.13.6.23

Changes for version 2.13.6.23 - 2026-07-14

  • add 'sjis' encoding, JISC Shift_JIS (JIS X 0201 and JIS X 0208 only, CP932 extended characters are not converted, use 'cp932' for them) with the eras 'sjis1978' (JIS C 6226-1978), 'sjis1983' (JIS X 0208-1983), 'sjis1990' (JIS X 0208-1990), 'sjis2000' (JIS X 0213:2000), and 'sjis2004' (JIS X 0213:2004). The year-less 'sjis' is the 1990 era. The eras differ by the JIS C 6226-1978 versus JIS X 0208-1983 exchanged/relocated codes, the two kanji appended by JIS X 0208-1990 (U+51DC, U+7199), and the ten kanji appended by JIS X 0213:2004.
  • add 'euc' encoding, JISC EUC-JP (code set 0, 1, 2, and 3), with the eras 'euc1978', 'euc1983', 'euc1990', 'euc2000', and 'euc2004' (the year-less 'euc' is the 1990 era). 'euc' of each era is derived by encoding conversion from the 'sjis' column of the same era via men-ku-ten, so no new mapping data is added for code set 0, 1, and 2
  • support EUC-JP code set 3 (SS3 0x8F + JIS X 0212) input/output on 'euc1990' (and so the year-less 'euc') for every character that has a row in the Jacode4e table through some OTHER encoding (CP932X, CP932, CP932IBM, CP932NEC, CP00930, KEIS78/83/90, JEF, JIPS(J)/(E), or LetsJ). A character whose ONLY repertoire membership is JIS X 0212 (about half of the JIS X 0212 repertoire; 3108 of 6067 code points in the source table) has no row and becomes GETA, because giving it one would require allocating a new UTF-8-SPUA-JP code point and is out of scope for this release; JIS X 0208 always wins over JIS X 0212 for a character that is in both (matches how 'sjis'/'euc'/'jis' behave for era differences). eras other than 1990 (1978, 1983, 2000, 2004) have no code set 3, since JIS X 0212 was specifically a 1990-era JIS supplement
  • the JIS X 0212 mapping is embedded as the "__EUC_X0212__" block appended after the main table in lib/Jacode4e.pm's __DATA__ section (JIS X 0212 GL octets + Unicode code point per line, derived from the JIS X 0212 (1990) to Unicode 1.1 Table already bundled at make__DATA__/LetsJ/...JIS0212.TXT); regenerate it with make__DATA__/JIS/fill_euc_codeset3.pl
  • add 'jis' encoding, JISC ISO-2022-JP (7bit JIS), with the eras 'jis1978', 'jis1983', 'jis1990', 'jis2000', and 'jis2004' (the year-less 'jis' is the 1990 era). 'jis' of each era shares the 'euc' table of the same era, so JIS X 0212 (ESC $ ( D) is available on 'jis1990' for the same character set as EUC-JP code set 3 of 'euc1990' (input is always accepted; output needs the JIS_X0212 option, see below) input escape sequences: ESC ( B, ESC ( J, ESC ( H, ESC ( I, ESC $ @, ESC $ B, ESC $ ( @, ESC $ ( B, ESC $ ( D, and ESC & @ ESC $ B 8bit JIS Katakana (GR octets 0xA1..0xDF) is also accepted on input SO (0x0E) and SI (0x0F) are input shift functions to JIS X 0201 Katakana (JIS7 katakana habit) with option JIS_SOSI => 1, and are data octets by default; SI restores the GL designation before SO output uses ESC ( B and ESC $ B with option OUTPUT_SHIFTING => 1; the DBCS designation on output is ESC $ @ for 'jis1978', ESC $ B for 'jis1983'/'jis1990', ESC $ ( O for 'jis2000', and ESC $ ( Q for 'jis2004'
  • add option JIS_X0212 => 1: a character that has no JIS X 0208 code but has a JIS X 0212 code is written as ESC $ ( D and JIS X 0212 codes on 'jis' output (ISO-2022-JP-1); requires OUTPUT_SHIFTING => 1 because escape sequences must be allowed in the output; default is false, effective only when $OUTPUT_encoding is 'jis'
  • add option JIS_KANA => 'GR' (default) / 'I' / 'SO': JIS X 0201 Katakana style on 'jis' output (GR octets, ESC ( I and GL octets, or SO + GL octets + SI)
  • add option JIS_DBCS => 'B' (default) / '@' / '&@B' / '(B' / '(@': DBCS escape sequence on 'jis' output
  • add option JIS_SBCS => 'B' (default) / 'J' / 'H': SBCS escape sequence on 'jis' output
  • add five columns "sjis1978", "sjis1983", "sjis1990", "sjis2000", and "sjis2004" to the conversion table in __DATA__ (row count is unchanged, all previously existing columns are unchanged), and make__DATA__/make__DATA__.pl generates them (the "euc" and "jis" columns are not stored in __DATA__; they are derived at load time from the "sjis" column of each era and the "__EUC_X0212__" block, see above)
  • add 'reset' hook to Jacode4e::convert() to initialize the kana shifting state of 'jis' encoding on every call
  • add doc/jacode4e_cheatsheet.*.txt (21 languages, same language set as the Jacode distribution: BM BN EN FR HI ID JA KM KO MN MY NE SI TH TL TR TW UR UZ VI ZH)
  • add t/0306_sjis_basic.t
  • add t/0307_euc_basic.t
  • add t/0308_jis_basic.t
  • add t/0309_sjis_euc_jis_roundtrip.t (whole code space roundtrip)
  • add t/0310_sjis_vs_cp932.t ('sjis' is JIS X 0201/0208 subset of 'cp932')
  • add t/0312_euc_codeset3.t (EUC-JP code set 3 / JIS X 0212 input/output and its coverage limits)
  • add t/0313_jis_variants_input.t
  • add t/0314_jis_variants_output.t
  • add t/0315_jis_x0212_output.t (JIS X 0212 output on 'jis' with the JIS_X0212 option)
  • add t/0318_sjis_euc_jis_era.t (era of 'sjis', 'euc', and 'jis')
  • add t/0319_euc_x0212_fill.t (repertoire coverage count, EUC code set 3 target uniqueness, JIS X 0208-over-JIS X 0212 precedence, and era isolation)
  • add t/9080_cheatsheets.t
  • add "sjis", "euc", and "jis" to encoding list of cheatsheet tests t/9001..t/9030
  • add the era encodings ('sjis1978'.. 'sjis2004', 'euc1978'.. 'euc2004', 'jis1978'.. 'jis2004') to --dumptable
  • package lib/Jacode4e.pm-2.13.6.23.TABLE.*.txt (generated by --dumptable; the EUC/JIS/SJIS dumps are one file per era, ...TABLE.EUC1978.txt .. ...TABLE.EUC2004.txt, etc.)
  • fix Jacode4e::convert() return value: a trailing shifting code (such as the last ESC ( B / ESC $ B of 'jis', SI/SO of 'cp00930' and 'keis*', etc.) was counted as one character; input consisting of shifting code(s) only now returns 0 (the closing shifting code for OUTPUT_SHIFTING is still emitted as before)
  • add t/0311_trailing_shift_count.t
  • integrate Jacode4e::RoundTrip into Jacode4e: add option ROUND_TRIP => 1 for round-trip (reversible) conversion. A character that has no native code in an encoding is allocated a code point from the user-defined (GAIJI) area of the encoding (cp932x, cp00930, keis78, keis83, keis90, jef, jef9p, jipsj, jipse, letsj, utf8, utf8.1, utf8jp). The default (no ROUND_TRIP option) keeps the ordinary Jacode4e behavior unchanged (a character with no native code becomes GETA). The ROUND_TRIP table is built lazily on the first ROUND_TRIP call (all data lines are kept in memory for this), so there is no extra cost when the option is not used. CAUTION: both "A to B" and "B to A" must specify ROUND_TRIP => 1 and use the same version of Jacode4e (the allocation order of the user-defined area depends on the data table of the version).
  • add --dumptable-roundtrip to dump the ROUND_TRIP table into "*.ROUNDTRIP.*.txt" files, separately from --dumptable
  • add t/0316_round_trip.t
  • add 'euc2004' encoding, JISC EUC-JIS-2004 (JIS X 0201, and JIS X 0213 plane 1 as two GR octets, plane 2 as SS3 0x8F + two GR octets). It is derived from the 'sjis2004' column via men-ku-ten (plane-row-cell), so no new mapping data is added.
  • add 'jis2004' encoding, JISC ISO-2022-JP-2004. input escape sequences: ESC ( B, ESC ( J, ESC ( I, ESC $ B, ESC $ @, ESC $ ( Q (JIS X 0213:2004 plane 1), ESC $ ( O (JIS X 0213:2000 plane 1), and ESC $ ( P (plane 2). output uses ESC $ ( Q for plane 1 (default), ESC $ ( P for plane 2, and ESC ( B for US-ASCII, with option OUTPUT_SHIFTING => 1.
  • add option JIS2004_PLANE1 => 'Q' (default, JIS X 0213:2004) or 'O' (JIS X 0213:2000): plane 1 designation escape sequence on 'jis2004' output
  • add 'euc2004' to --dumptable (the euc2004 table dumps plane 1 rows and plane 2 rows in EUC-JIS-2004 code set 3 form, three-octet row labels such as "EUC2004 8FA1A0")
  • add t/0317_euc2004_jis2004.t
  • restore the full-width "|" guide notation in the __DATA__ header (one "|" per code digit of each already-defined column, matching the "#1234567 1234 ..." ruler line); the guide lines are comments (skipped by the "next if /^#/" parser), so this is cosmetic only
  • regenerate doc/jacode4e_cheatsheet.*.txt (all 21 languages) with: the year editions of 'sjis'/'euc'/'jis' (1978/1983/1990/2000/2004, the year-less name is the 1990 edition), a beginner-friendly section on the differences between the JIS editions by year, all twelve { %option } keys each with a description and a call example (the ROUND_TRIP option is documented in the same style as the others, not as a separate caution), and runnable programming samples
  • add doc/build_cheatsheets.pl to generate the 21 cheat sheets from one shared template (Perl 5.005_03 compatible)
  • created by INABA Hitoshi

Documentation

SECURITY.md

Modules

Jacode4e
Converts Character Encodings for Enterprise in Japan

Other files