aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
Diffstat
-rw-r--r--ChangeLog11
-rw-r--r--transcode.c84
-rw-r--r--transcode_data.h39
-rw-r--r--transcode_data_japanese.c207
4 files changed, 322 insertions, 19 deletions
diff --git a/ChangeLog b/ChangeLog
index 413eaa5288..ba3ba8c020 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,14 @@
+Mon Dec 24 22:46:42 2007 NARUSE, Yui <naruse@ruby-lang.org>
+
+ * transocode.c: register_functional_transcoder() added.
+ (init_transcoder_table(: register ISO-2022-JP.
+ (str_transcode): add preprocessor and postprocessor.
+
+ * transcode_data_japanese.c: add ISO-2022-JP support.
+
+ * transcode_data.h: moved transcoder and transcoding difinition from
+ transcode.c.
+
Mon Dec 24 20:29:28 2007 Koichi Sasada <ko1@atdot.net>
* test/io/nonblock/test_flush.rb: fix test for 1.9.
diff --git a/transcode.c b/transcode.c
index 82daf5b929..281ecba697 100644
--- a/transcode.c
+++ b/transcode.c
@@ -60,22 +60,17 @@ extern const BYTE_LOOKUP rb_from_EUC_JP;
extern const BYTE_LOOKUP rb_to_SHIFT_JIS;
extern const BYTE_LOOKUP rb_to_EUC_JP;
+extern void from_iso_2022_jp_transcoder_preprocessor(char**, char**, char*, char*,
+ struct transcoder_st *transcoder, struct transcoding*);
+extern void to_iso_2022_jp_transcoder_postprocessor(char**, char**, char*, char*,
+ struct transcoder_st *transcoder, struct transcoding*);
/* declarations probably need to go into separate header file, e.g. transcode.h */
-/* static structure, one per supported encoding pair */
-typedef struct {
- const char *from_encoding;
- const char *to_encoding;
- const BYTE_LOOKUP *conv_tree_start;
- int max_output;
- int from_utf8;
-} transcoder;
-
/* todo: dynamic structure, one per conversion (stream) */
/* in the future, add some mechanism for dynamically adding stuff here */
-#define MAX_TRANSCODERS 33 /* todo: fix: this number has to be adjusted by hand */
+#define MAX_TRANSCODERS 35 /* todo: fix: this number has to be adjusted by hand */
static transcoder transcoder_table[MAX_TRANSCODERS];
/* not sure why it's not possible to do relocatable initializations */
@@ -100,6 +95,29 @@ register_transcoder(const char *from_e, const char *to_e,
}
static void
+register_functional_transcoder(const char *from_e, const char *to_e,
+ const BYTE_LOOKUP *tree_start, int max_output, int from_utf8,
+ void (*preprocessor)(char**, char**, char*, char*, transcoder*, transcoding*),
+ void (*postprocessor)(char**, char**, char*, char*, transcoder*, transcoding*))
+{
+ static int n = 0;
+ if (n >= MAX_TRANSCODERS) {
+ /* we are initializing, is it okay to use rb_raise here? */
+ rb_raise(rb_eRuntimeError /*change exception*/, "not enough transcoder slots");
+ }
+ transcoder_table[n].from_encoding = from_e;
+ transcoder_table[n].to_encoding = to_e;
+ transcoder_table[n].conv_tree_start = tree_start;
+ transcoder_table[n].max_output = max_output;
+ transcoder_table[n].from_utf8 = from_utf8;
+ transcoder_table[n].conv_tree_start = tree_start;
+ transcoder_table[n].preprocessor = preprocessor;
+ transcoder_table[n].postprocessor = postprocessor;
+
+ n++;
+}
+
+static void
init_transcoder_table(void)
{
register_transcoder("ISO-8859-1", "UTF-8", &rb_from_ISO_8859_1, 2, 0);
@@ -135,6 +153,10 @@ init_transcoder_table(void)
register_transcoder("EUC-JP", "UTF-8", &rb_from_EUC_JP, 3, 0);
register_transcoder("UTF-8", "SHIFT_JIS", &rb_to_SHIFT_JIS, 2, 1);
register_transcoder("UTF-8", "EUC-JP", &rb_to_EUC_JP, 2, 1);
+ register_functional_transcoder("ISO-2022-JP", "UTF-8", &rb_from_EUC_JP,
+ 8, 0, &from_iso_2022_jp_transcoder_preprocessor, NULL);
+ register_functional_transcoder("UTF-8", "ISO-2022-JP", &rb_to_EUC_JP,
+ 8, 1, NULL, &to_iso_2022_jp_transcoder_postprocessor);
register_transcoder(NULL, NULL, NULL, 0, 0);
}
@@ -165,14 +187,6 @@ transcode_dispatch(const char* from_encoding, const char* to_encoding)
return NULL;
}
-/* dynamic structure, one per conversion (similar to iconv_t) */
-/* may carry conversion state (e.g. for iso-2022-jp) */
-typedef struct transcoding {
- VALUE ruby_string_dest; /* the String used as the conversion destination,
- or NULL if something else is being converted */
- char *(*flush_func)(struct transcoding*, int, int);
-} transcoding;
-
/*
* Transcoding engine logic
@@ -331,6 +345,23 @@ str_transcode(int argc, VALUE *argv, VALUE *self)
rb_raise(rb_eArgError, "transcoding not supported (from %s to %s)", from_e, to_e);
}
+ if (my_transcoder->preprocessor)
+ {
+ fromp = sp = RSTRING_PTR(str);
+ slen = RSTRING_LEN(str);
+ blen = slen + 30; /* len + margin */
+ dest = rb_str_tmp_new(blen);
+ bp = RSTRING_PTR(dest);
+ my_transcoding.ruby_string_dest = dest;
+ (*my_transcoder->preprocessor)(&fromp, &bp, (sp+slen), (bp+blen), my_transcoder, &my_transcoding);
+ if (fromp != sp+slen) {
+ rb_raise(rb_eArgError, "not fully converted, %d bytes left", sp+slen-fromp);
+ }
+ buf = RSTRING_PTR(dest);
+ *bp = '\0';
+ rb_str_set_len(dest, bp - buf);
+ str = dest;
+ }
fromp = sp = RSTRING_PTR(str);
slen = RSTRING_LEN(str);
blen = slen + 30; /* len + margin */
@@ -346,6 +377,23 @@ str_transcode(int argc, VALUE *argv, VALUE *self)
buf = RSTRING_PTR(dest);
*bp = '\0';
rb_str_set_len(dest, bp - buf);
+ if (my_transcoder->postprocessor)
+ {
+ str = dest;
+ fromp = sp = RSTRING_PTR(str);
+ slen = RSTRING_LEN(str);
+ blen = slen + 30; /* len + margin */
+ dest = rb_str_tmp_new(blen);
+ bp = RSTRING_PTR(dest);
+ my_transcoding.ruby_string_dest = dest;
+ (*my_transcoder->postprocessor)(&fromp, &bp, (sp+slen), (bp+blen), my_transcoder, &my_transcoding);
+ if (fromp != sp+slen) {
+ rb_raise(rb_eArgError, "not fully converted, %d bytes left", sp+slen-fromp);
+ }
+ buf = RSTRING_PTR(dest);
+ *bp = '\0';
+ rb_str_set_len(dest, bp - buf);
+ }
if (encoding_equal(my_transcoder->to_encoding, to_e)) {
final_encoding = 1;
diff --git a/transcode_data.h b/transcode_data.h
index dc90c92a51..3131877a41 100644
--- a/transcode_data.h
+++ b/transcode_data.h
@@ -1,3 +1,20 @@
+/**********************************************************************
+
+ transcode_data.h -
+
+ $Author$
+ $Date$
+ created at: Mon 10 Dec 2007 14:01:47 JST 2007
+
+ Copyright (C) 2007 Martin Duerst
+
+**********************************************************************/
+
+#include "ruby/ruby.h"
+
+#ifndef RUBY_TRANSCODE_DATA_H
+#define RUBY_TRANSCODE_DATA_H 1
+
typedef unsigned char base_element;
typedef struct byte_lookup {
@@ -37,3 +54,25 @@ typedef struct byte_lookup {
#define TWOTRAIL /* legal but undefined if two more trailing UTF-8 */
#define THREETRAIL /* legal but undefined if three more trailing UTF-8 */
+/* dynamic structure, one per conversion (similar to iconv_t) */
+/* may carry conversion state (e.g. for iso-2022-jp) */
+typedef struct transcoding {
+ VALUE ruby_string_dest; /* the String used as the conversion destination,
+ or NULL if something else is being converted */
+ char *(*flush_func)(struct transcoding*, int, int);
+} transcoding;
+
+/* static structure, one per supported encoding pair */
+typedef struct transcoder_st{
+ const char *from_encoding;
+ const char *to_encoding;
+ const BYTE_LOOKUP *conv_tree_start;
+ int max_output;
+ int from_utf8;
+ void (*preprocessor)(char**, char**, char*, char*,
+ struct transcoder_st *transcoder, struct transcoding*);
+ void (*postprocessor)(char**, char**, char*, char*,
+ struct transcoder_st *transcoder, struct transcoding*);
+} transcoder;
+
+#endif /* RUBY_TRANSCODE_DATA_H */
diff --git a/transcode_data_japanese.c b/transcode_data_japanese.c
index fc7243186d..223d1586e3 100644
--- a/transcode_data_japanese.c
+++ b/transcode_data_japanese.c
@@ -23618,4 +23618,209 @@ rb_to_EUC_JP = {
to_EUC_JP_infos
};
-/* Footprint (bytes): gross: 212680, saved: 50764, net: 161916 */
+#define ISO_2022_ENCODING(escseq, byte) ((escseq<<8)|byte)
+enum ISO_2022_ESCSEQ {
+ ISO_2022_CZD = '!',
+ ISO_2022_C1D = '"',
+ ISO_2022_GZD4 = '(',
+ ISO_2022_G1D4 = ')',
+ ISO_2022_G2D4 = '*',
+ ISO_2022_G3D4 = '+',
+ ISO_2022_G1D6 = '-',
+ ISO_2022_G2D6 = '.',
+ ISO_2022_G3D6 = '/',
+ ISO_2022_GZDM4 = ISO_2022_ENCODING('$','('),
+ ISO_2022_G1DM4 = ISO_2022_ENCODING('$',')'),
+ ISO_2022_G2DM4 = ISO_2022_ENCODING('$','*'),
+ ISO_2022_G3DM4 = ISO_2022_ENCODING('$','+'),
+ ISO_2022_G1DM6 = ISO_2022_ENCODING('$','-'),
+ ISO_2022_G2DM6 = ISO_2022_ENCODING('$','.'),
+ ISO_2022_G3DM6 = ISO_2022_ENCODING('$','/'),
+ ISO_2022_DOCS = ISO_2022_ENCODING('%','I'),
+ ISO_2022_IRR = '&'
+};
+
+
+#define ISO_2022_GZ_ASCII ISO_2022_ENCODING(ISO_2022_GZD4, 'B')
+#define ISO_2022_GZ_JIS_X_0201_Katakana ISO_2022_ENCODING(ISO_2022_GZD4, 'I')
+#define ISO_2022_GZ_JIS_X_0201_Roman ISO_2022_ENCODING(ISO_2022_GZD4, 'J')
+#define ISO_2022_GZ_JIS_C_6226_1978 ISO_2022_ENCODING(ISO_2022_GZDM4,'@')
+#define ISO_2022_GZ_JIS_X_0208_1983 ISO_2022_ENCODING(ISO_2022_GZDM4,'B')
+#define ISO_2022_GZ_JIS_X_0212_1990 ISO_2022_ENCODING(ISO_2022_GZDM4,'D')
+#define ISO_2022_GZ_JIS_X_0213_2000_1 ISO_2022_ENCODING(ISO_2022_GZDM4,'O')
+#define ISO_2022_GZ_JIS_X_0213_2000_2 ISO_2022_ENCODING(ISO_2022_GZDM4,'P')
+#define ISO_2022_GZ_JIS_X_0213_2004_1 ISO_2022_ENCODING(ISO_2022_GZDM4,'Q')
+
+static int
+get_iso_2022_mode(char **in_pos)
+{
+ int new_mode;
+ char *in_p = *in_pos;
+ switch (*in_p++)
+ {
+ case '(':
+ switch (*in_p++)
+ {
+ case 'B': case 'I': case 'J':
+ new_mode = ISO_2022_ENCODING(ISO_2022_GZD4, *(in_p-1));
+ break;
+ default:
+ rb_raise(rb_eRuntimeError /*change exception*/, "this mode is not supported (ESC ( %c)", *(in_p-1));
+ break;
+ }
+ break;
+ case '$':
+ switch (*in_p++)
+ {
+ case '@': case 'A': case 'B':
+ new_mode = ISO_2022_ENCODING(ISO_2022_GZDM4, *(in_p-1));
+ break;
+ case '(':
+ switch (*in_p++)
+ {
+ case 'D': case 'O': case 'P': case 'Q':
+ new_mode = ISO_2022_ENCODING(ISO_2022_GZDM4, *(in_p-1));
+ break;
+ default:
+ rb_raise(rb_eRuntimeError /*change exception*/, "this mode is not supported (ESC $ ( %c)", *(in_p-1));
+ break;
+ }
+ break;
+ default:
+ rb_raise(rb_eRuntimeError /*change exception*/, "this mode is not supported (ESC $ %c)", *(in_p-1));
+ break;
+ }
+ break;
+ default:
+ rb_raise(rb_eRuntimeError /*change exception*/, "this mode is not supported (ESC %c)", *(in_p-1));
+ break;
+ }
+ *in_pos = in_p;
+ return new_mode;
+}
+
+void
+from_iso_2022_jp_transcoder_preprocessor(char **in_pos, char **out_pos,
+ char *in_stop, char *out_stop,
+ transcoder *my_transcoder,
+ transcoding *my_transcoding)
+{
+ char *in_p = *in_pos, *out_p = *out_pos;
+ int cur_mode = ISO_2022_GZ_ASCII;
+ unsigned char c1;
+ char *out_s = out_stop - my_transcoder->max_output + 1;
+ while (in_p < in_stop) {
+ if (out_p >= out_s) {
+ int len = (out_p - *out_pos);
+ int new_len = (len + my_transcoder->max_output) * 2;
+ *out_pos = (*my_transcoding->flush_func)(my_transcoding, len, new_len);
+ out_p = *out_pos + len;
+ out_s = *out_pos + new_len - my_transcoder->max_output;
+ }
+ c1 = *in_p++;
+ if (c1 == 0x1B) {
+ cur_mode = get_iso_2022_mode(&in_p);
+ } else if (c1 == 0x1E || c1 == 0x1F) {
+ /* SHIFT */
+ rb_raise(rb_eRuntimeError /*change exception*/, "shift is not supported");
+ } else if (c1 >= 0x80) {
+ rb_raise(rb_eRuntimeError /*change exception*/, "illegal byte sequence");
+ } else {
+ switch (cur_mode) {
+ case ISO_2022_GZ_ASCII:
+ case ISO_2022_GZ_JIS_X_0201_Roman:
+ *out_p++ = c1;
+ break;
+ case ISO_2022_GZ_JIS_X_0201_Katakana:
+ *out_p++ = 0x8E;
+ *out_p++ = c1 | 0x80;
+ break;
+ case ISO_2022_GZ_JIS_X_0212_1990:
+ *out_p++ = 0x8F;
+ case ISO_2022_GZ_JIS_C_6226_1978:
+ case ISO_2022_GZ_JIS_X_0208_1983:
+ *out_p++ = c1 | 0x80;
+ *out_p++ = *in_p++ | 0x80;
+ break;
+ }
+ }
+ }
+ /* cleanup */
+ *in_pos = in_p;
+ *out_pos = out_p;
+}
+
+static int
+select_iso_2022_mode(char **out_pos, int new_mode)
+{
+ char *out_p = *out_pos;
+ *out_p++ = '\e';
+ switch (new_mode>>8)
+ {
+ case ISO_2022_GZD4:
+ *out_p++ = new_mode >> 8;
+ *out_p++ = new_mode & 0x7F;
+ break;
+ case ISO_2022_GZDM4:
+ *out_p++ = new_mode >> 16;
+ if ((new_mode & 0x7F) != '@' &&
+ (new_mode & 0x7F) != 'A' &&
+ (new_mode & 0x7F) != 'B')
+ {
+ *out_p++ = (new_mode>>8) & 0x7F;
+ }
+ *out_p++ = new_mode & 0x7F;
+ break;
+ default:
+ rb_raise(rb_eRuntimeError /*change exception*/, "this mode is not supported.");
+ break;
+ }
+ *out_pos = out_p;
+ return new_mode;
+}
+
+void
+to_iso_2022_jp_transcoder_postprocessor(char **in_pos, char **out_pos,
+ char *in_stop, char *out_stop,
+ transcoder *my_transcoder,
+ transcoding *my_transcoding)
+{
+ char *in_p = *in_pos, *out_p = *out_pos;
+ int cur_mode = ISO_2022_GZ_ASCII, new_mode = 0;
+ unsigned char next_byte;
+ char *out_s = out_stop - my_transcoder->max_output + 1;
+ while (in_p < in_stop) {
+ if (out_p >= out_s) {
+ int len = (out_p - *out_pos);
+ int new_len = (len + my_transcoder->max_output) * 2;
+ *out_pos = (*my_transcoding->flush_func)(my_transcoding, len, new_len);
+ out_p = *out_pos + len;
+ out_s = *out_pos + new_len - my_transcoder->max_output;
+ }
+ next_byte = *in_p++;
+ if (next_byte < 0x80) {
+ new_mode = ISO_2022_GZ_ASCII;
+ } else if (next_byte == 0x8E) {
+ new_mode = ISO_2022_GZ_JIS_X_0201_Katakana;
+ next_byte = *in_p++;
+ } else if (next_byte == 0x8F) {
+ new_mode = ISO_2022_GZ_JIS_X_0212_1990;
+ next_byte = *in_p++;
+ } else {
+ new_mode = ISO_2022_GZ_JIS_X_0208_1983;
+ }
+ if (cur_mode != new_mode)
+ cur_mode = select_iso_2022_mode(&out_p, new_mode);
+ if (cur_mode < 0xFFFF) {
+ *out_p++ = next_byte & 0x7F;
+ } else {
+ *out_p++ = next_byte & 0x7F;
+ *out_p++ = *in_p++ & 0x7F;
+ }
+ }
+ if (cur_mode != ISO_2022_GZ_ASCII)
+ cur_mode = select_iso_2022_mode(&out_p, ISO_2022_GZ_ASCII);
+ /* cleanup */
+ *in_pos = in_p;
+ *out_pos = out_p;
+}
generated by cgit v1.2.3 (git 2.39.1) at 2024-06-23 14:05:37 +0000