/* Header for code conversion stuff
Copyright (C) 1991, 1995 Free Software Foundation, Inc.
Copyright (C) 1995 Sun Microsystems, Inc.
+ Copyright (C) 1999,2000,2002,2005,2012 MORIOKA Tomohiko
This file is part of XEmacs.
/* 91.10.09 written by K.Handa <handa@etl.go.jp> */
/* Rewritten by Ben Wing <ben@xemacs.org>. */
-#ifndef _XEMACS_MULE_CODING_H_
-#define _XEMACS_MULE_CODING_H_
+#ifndef INCLUDED_file_coding_h_
+#define INCLUDED_file_coding_h_
struct decoding_stream;
struct encoding_stream;
/* Coding system types. These go into the TYPE field of a
- struct Lisp_Coding_System. */
+ Lisp_Coding_System. */
enum coding_system_type
{
CODESYS_BIG5, /* BIG5 (used for Taiwanese). */
CODESYS_UCS4, /* ISO 10646 UCS-4 */
CODESYS_UTF8, /* ISO 10646 UTF-8 */
+#ifdef UTF2000
+ CODESYS_UTF16, /* ISO 10646 UTF-16 */
+#endif
CODESYS_CCL, /* Converter written in CCL. */
#endif
CODESYS_NO_CONVERSION /* "No conversion"; used for binary files.
struct lcrecord_header header;
/* Name and doc string of this coding system. */
- Lisp_Object name, doc_string;
+ Lisp_Object name;
+ Lisp_Object doc_string;
/* This is the major type of the coding system -- one of Big5, ISO2022,
Shift-JIS, etc. See the constants above. */
system is active for a particular buffer. */
Lisp_Object mnemonic;
- Lisp_Object post_read_conversion, pre_write_conversion;
+ Lisp_Object post_read_conversion;
+ Lisp_Object pre_write_conversion;
eol_type_t eol_type;
/* Subsidiary coding systems that specify a particular type of EOL
marking, rather than autodetecting it. These will only be non-nil
if (eol_type == EOL_AUTODETECT). */
- Lisp_Object eol_lf, eol_crlf, eol_cr;
+ Lisp_Object eol_lf;
+ Lisp_Object eol_crlf;
+ Lisp_Object eol_cr;
#ifdef MULE
struct
{
{
/* For a CCL coding system, these specify the CCL programs used for
decoding (input) and encoding (output). */
- Lisp_Object decode, encode;
+ Lisp_Object decode;
+ Lisp_Object encode;
} ccl;
#endif
+#ifdef UTF2000
+ unsigned int disable_composition :1;
+ unsigned int enable_decomposition :1;
+ unsigned int use_entity_reference :1;
+ Lisp_Object ccs_priority_list;
+#endif
};
typedef struct Lisp_Coding_System Lisp_Coding_System;
-DECLARE_LRECORD (coding_system, struct Lisp_Coding_System);
-#define XCODING_SYSTEM(x) XRECORD (x, coding_system, struct Lisp_Coding_System)
+DECLARE_LRECORD (coding_system, Lisp_Coding_System);
+#define XCODING_SYSTEM(x) XRECORD (x, coding_system, Lisp_Coding_System)
#define XSETCODING_SYSTEM(x, p) XSETRECORD (x, p, coding_system)
#define CODING_SYSTEMP(x) RECORDP (x, coding_system)
#define CHECK_CODING_SYSTEM(x) CHECK_RECORD (x, coding_system)
#define CODING_SYSTEM_CCL_DECODE(codesys) ((codesys)->ccl.decode)
#define CODING_SYSTEM_CCL_ENCODE(codesys) ((codesys)->ccl.encode)
#endif /* MULE */
+#ifdef UTF2000
+#define CODING_SYSTEM_DISABLE_COMPOSITION(codesys) \
+ ((codesys)->disable_composition)
+#define CODING_SYSTEM_ENABLE_DECOMPOSITION(codesys) \
+ ((codesys)->enable_decomposition)
+#define CODING_SYSTEM_USE_ENTITY_REFERENCE(codesys) \
+ ((codesys)->use_entity_reference)
+#define CODING_SYSTEM_CCS_PRIORITY_LIST(codesys) \
+ ((codesys)->ccs_priority_list)
+#endif
#define XCODING_SYSTEM_NAME(codesys) \
CODING_SYSTEM_NAME (XCODING_SYSTEM (codesys))
#define XCODING_SYSTEM_CCL_ENCODE(codesys) \
CODING_SYSTEM_CCL_ENCODE (XCODING_SYSTEM (codesys))
#endif /* MULE */
+#ifdef UTF2000
+#define XCODING_SYSTEM_DISABLE_COMPOSITION(codesys) \
+ CODING_SYSTEM_DISABLE_COMPOSITION (XCODING_SYSTEM (codesys))
+#define XCODING_SYSTEM_ENABLE_DECOMPOSITION(codesys) \
+ CODING_SYSTEM_ENABLE_DECOMPOSITION (XCODING_SYSTEM (codesys))
+#define XCODING_SYSTEM_USE_ENTITY_REFERENCE(codesys) \
+ CODING_SYSTEM_USE_ENTITY_REFERENCE (XCODING_SYSTEM (codesys))
+#define XCODING_SYSTEM_CCS_PRIORITY_LIST(codesys) \
+ CODING_SYSTEM_CCS_PRIORITY_LIST (XCODING_SYSTEM (codesys))
+#endif
EXFUN (Fcoding_category_list, 0);
EXFUN (Fcoding_category_system, 1);
extern Lisp_Object Qucs4, Qutf8;
extern Lisp_Object Qbig5, Qccl, Qcharset_g0;
extern Lisp_Object Qcharset_g1, Qcharset_g2, Qcharset_g3, Qcoding_system_error;
-extern Lisp_Object Qcoding_systemp, Qcr, Qcrlf, Qctext, Qdecode, Qencode;
+extern Lisp_Object Qcoding_systemp, Qcr, Qcrlf, Qdecode, Qencode;
extern Lisp_Object Qeol_cr, Qeol_crlf, Qeol_lf, Qeol_type, Qescape_quoted;
extern Lisp_Object Qforce_g0_on_output, Qforce_g1_on_output;
extern Lisp_Object Qforce_g2_on_output, Qforce_g3_on_output;
#define ISO_CODE_CSI 0x9B /* control-sequence-introduce */
#endif /* MULE */
-/* For detecting the encoding of text */
+/* Distinguishable categories of encodings.
+
+ This list determines the initial priority of the categories.
+
+ For better or worse, currently Mule files are encoded in 7-bit ISO 2022.
+ For this reason, under Mule ISO_7 gets highest priority.
+
+ Putting NO_CONVERSION second prevents "binary corruption" in the
+ default case in all but the (presumably) extremely rare case of a
+ binary file which contains redundant escape sequences but no 8-bit
+ characters.
+
+ The remaining priorities are based on perceived "internationalization
+ political correctness." An exception is UCS-4 at the bottom, since
+ basically everything is compatible with UCS-4, but it is likely to
+ be very rare as an external encoding. */
+
enum coding_category_type
{
+ /* must be a contiguous range of values 0 -- CODING_CATEGORY_LAST - 1 */
#ifdef MULE
- CODING_CATEGORY_SHIFT_JIS,
CODING_CATEGORY_ISO_7, /* ISO2022 system using only seven-bit bytes,
no locking shift */
- CODING_CATEGORY_ISO_8_DESIGNATE, /* ISO2022 system using eight-bit bytes,
- no locking shift, no single shift,
- using designation to switch charsets */
+ CODING_CATEGORY_NO_CONVERSION,
+ CODING_CATEGORY_UTF8,
CODING_CATEGORY_ISO_8_1, /* ISO2022 system using eight-bit bytes,
no locking shift, no designation sequences,
one-dimension characters in the upper half. */
CODING_CATEGORY_ISO_8_2, /* ISO2022 system using eight-bit bytes,
no locking shift, no designation sequences,
two-dimension characters in the upper half. */
+ CODING_CATEGORY_ISO_8_DESIGNATE, /* ISO2022 system using eight-bit bytes,
+ no locking shift, no single shift,
+ using designation to switch charsets */
CODING_CATEGORY_ISO_LOCK_SHIFT, /* ISO2022 system using locking shift */
+ CODING_CATEGORY_SHIFT_JIS,
CODING_CATEGORY_BIG5,
+ CODING_CATEGORY_UTF16,
CODING_CATEGORY_UCS4,
- CODING_CATEGORY_UTF8,
+#else /* not MULE */
+ CODING_CATEGORY_NO_CONVERSION,
#endif /* MULE */
- CODING_CATEGORY_NO_CONVERSION
+ CODING_CATEGORY_LAST /* not a real coding category */
};
-#define CODING_CATEGORY_LAST CODING_CATEGORY_NO_CONVERSION
-
#ifdef MULE
#define CODING_CATEGORY_SHIFT_JIS_MASK \
(1 << CODING_CATEGORY_SHIFT_JIS)
(1 << CODING_CATEGORY_BIG5)
#define CODING_CATEGORY_UCS4_MASK \
(1 << CODING_CATEGORY_UCS4)
+#define CODING_CATEGORY_UTF16_MASK \
+ (1 << CODING_CATEGORY_UTF16)
#define CODING_CATEGORY_UTF8_MASK \
(1 << CODING_CATEGORY_UTF8)
#endif
#define BUFBYTE_FIRST_BYTE_P(c) ((c) < 0xA0)
#define BUFBYTE_LEADING_BYTE_P(c) BYTE_C1_P (c)
#endif /* not MULE */
-#endif /* _XEMACS_MULE_CODING_H_ */
+
+#endif /* INCLUDED_file_coding_h_ */
projects / chise / xemacs-chise.git.1 / blobdiff