Moriyoshi Koizumi
moriy****@users*****
2004年 2月 4日 (水) 14:56:44 JST
moriyoshi 04/02/04 14:56:44
Modified: mbfl Tag: LIBMBFL_1_0 mbfilter.c mbfl_convert.c
mbfl_convert.h mbfl_ident.c mbfl_ident.h
Log:
- MFH: No need to export mbfl_*_select_vtbl()
- MFH: Thoroughfully overhaul mbfl_identify_encoding()
Revision Changes Path
No revision
No revision
1.8.2.6 +27 -42 libmbfl/mbfl/mbfilter.c
Index: mbfilter.c
===================================================================
RCS file: /cvsroot/php-i18n/libmbfl/mbfl/mbfilter.c,v
retrieving revision 1.8.2.5
retrieving revision 1.8.2.6
diff -u -r1.8.2.5 -r1.8.2.6
--- mbfilter.c 4 Aug 2003 23:11:42 -0000 1.8.2.5
+++ mbfilter.c 4 Feb 2004 05:56:44 -0000 1.8.2.6
@@ -334,14 +334,14 @@
* encoding detector
*/
mbfl_encoding_detector *
-mbfl_encoding_detector_new(enum mbfl_no_encoding *elist, int eliztsz)
+mbfl_encoding_detector_new(enum mbfl_no_encoding *elist, int elistsz)
{
mbfl_encoding_detector *identd;
int i, num;
mbfl_identify_filter *filter;
- if (elist == NULL || eliztsz <= 0) {
+ if (elist == NULL || elistsz <= 0) {
return NULL;
}
@@ -350,7 +350,7 @@
if (identd == NULL) {
return NULL;
}
- identd->filter_list = (mbfl_identify_filter **)mbfl_calloc(eliztsz, sizeof(mbfl_identify_filter *));
+ identd->filter_list = (mbfl_identify_filter **)mbfl_calloc(elistsz, sizeof(mbfl_identify_filter *));
if (identd->filter_list == NULL) {
mbfl_free(identd);
return NULL;
@@ -359,7 +359,7 @@
/* create filters */
i = 0;
num = 0;
- while (i < eliztsz) {
+ while (i < elistsz) {
filter = mbfl_identify_filter_new(elist[i]);
if (filter != NULL) {
identd->filter_list[num] = filter;
@@ -522,43 +522,36 @@
* identify encoding
*/
const mbfl_encoding *
-mbfl_identify_encoding(mbfl_string *string, enum mbfl_no_encoding *elist, int eliztsz, int strict)
+mbfl_identify_encoding(mbfl_string *string, enum mbfl_no_encoding *elist, int elistsz, int strict)
{
int i, n, num, bad;
unsigned char *p;
- const struct mbfl_identify_vtbl *vtbl;
mbfl_identify_filter *flist, *filter;
const mbfl_encoding *encoding;
- /* initialize */
- flist = (mbfl_identify_filter *)mbfl_calloc(eliztsz, sizeof(mbfl_identify_filter));
+ /* flist is an array of mbfl_identify_filter instances */
+ flist = (mbfl_identify_filter *)mbfl_calloc(elistsz, sizeof(mbfl_identify_filter));
if (flist == NULL) {
return NULL;
}
- i = 0;
+
num = 0;
if (elist != NULL) {
- while (i < eliztsz) {
- vtbl = mbfl_identify_filter_get_vtbl(elist[i]);
- if (vtbl != NULL) {
- filter = &flist[num];
- mbfl_identify_filter_set_vtbl(filter, vtbl);
- filter->encoding = mbfl_no2encoding(vtbl->encoding);
- (*filter->filter_ctor)(filter);
+ for (i = 0; i < elistsz; i++) {
+ if (!mbfl_identify_filter_init(&flist[num], elist[i])) {
num++;
}
- i++;
}
}
/* feed data */
n = string->len;
p = string->val;
+
if (p != NULL) {
+ bad = 0;
while (n > 0) {
- i = 0;
- bad = 0;
- while (i < num) {
+ for (i = 0; i < num; i++) {
filter = &flist[i];
(*filter->filter_function)(*p, filter);
if (filter->flag) {
@@ -566,7 +559,7 @@
}
i++;
}
- if ((num - 1) <= bad && !strict) {
+ if ((num - 1) <= bad) {
break;
}
p++;
@@ -575,41 +568,33 @@
}
/* judge */
- i = num - 1;
- bad = 1;
encoding = NULL;
- while (i >= 0) {
+
+ for (i = 0; i < num; i++) {
filter = &flist[i];
- if (filter->flag) {
- bad++;
- } else {
+ if (!filter->flag) {
encoding = filter->encoding;
+ break;
}
- i--;
}
-#if 0
- if (bad < num) {
- encoding = NULL;
- }
-#endif
- i = 0;
- while (i < num) {
- filter = &flist[i];
- (*filter->filter_dtor)(filter);
- i++;
+ /* cleanup */
+ /* dtors should be called in reverse order */
+ i = num; while (--i >= 0) {
+ mbfl_identify_filter_cleanup(&flist[i]);
}
+
mbfl_free((void *)flist);
return encoding;
}
const char*
-mbfl_identify_encoding_name(mbfl_string *string, enum mbfl_no_encoding *elist, int eliztsz, int strict)
+mbfl_identify_encoding_name(mbfl_string *string, enum mbfl_no_encoding *elist, int elistsz, int strict)
{
const mbfl_encoding *encoding;
- encoding = mbfl_identify_encoding(string, elist, eliztsz, strict);
+ encoding = mbfl_identify_encoding(string, elist, elistsz, strict);
if (encoding != NULL &&
encoding->no_encoding > mbfl_no_encoding_charset_min &&
encoding->no_encoding < mbfl_no_encoding_charset_max) {
@@ -620,11 +605,11 @@
}
const enum mbfl_no_encoding
-mbfl_identify_encoding_no(mbfl_string *string, enum mbfl_no_encoding *elist, int eliztsz)
+mbfl_identify_encoding_no(mbfl_string *string, enum mbfl_no_encoding *elist, int elistsz)
{
const mbfl_encoding *encoding;
- encoding = mbfl_identify_encoding(string, elist, eliztsz, 0);
+ encoding = mbfl_identify_encoding(string, elist, elistsz, 0);
if (encoding != NULL &&
encoding->no_encoding > mbfl_no_encoding_charset_min &&
encoding->no_encoding < mbfl_no_encoding_charset_max) {
1.1.2.3 +7 -8 libmbfl/mbfl/mbfl_convert.c
Index: mbfl_convert.c
===================================================================
RCS file: /cvsroot/php-i18n/libmbfl/mbfl/mbfl_convert.c,v
retrieving revision 1.1.2.2
retrieving revision 1.1.2.3
diff -u -r1.1.2.2 -r1.1.2.3
--- mbfl_convert.c 4 Feb 2004 01:58:09 -0000 1.1.2.2
+++ mbfl_convert.c 4 Feb 2004 05:56:44 -0000 1.1.2.3
@@ -89,6 +89,8 @@
#include "filters/mbfilter_ucs2.h"
#include "filters/mbfilter_htmlent.h"
+static void mbfl_convert_filter_reset_vtbl(mbfl_convert_filter *filter);
+
/* hex character table "0123456789ABCDEF" */
static char mbfl_hexchar_table[] = {
0x30,0x31,0x32,0x33,0x34,0x35,0x36,0x37,0x38,0x39,0x41,0x42,0x43,0x44,0x45,0x46
@@ -247,7 +249,7 @@
filter->illegal_substchar = 0x3f; /* '?' */
/* setup the function table */
- mbfl_convert_filter_select_vtbl(filter);
+ mbfl_convert_filter_reset_vtbl(filter);
/* constructor */
(*filter->filter_ctor)(filter);
@@ -277,11 +279,8 @@
return (filter->flush_function ? (*filter->flush_function)(filter->data) : 0);
}
-void
-mbfl_convert_filter_reset(
- mbfl_convert_filter *filter,
- enum mbfl_no_encoding from,
- enum mbfl_no_encoding to)
+void mbfl_convert_filter_reset(mbfl_convert_filter *filter,
+ enum mbfl_no_encoding from, enum mbfl_no_encoding to)
{
/* destruct old filter */
(*filter->filter_dtor)(filter);
@@ -291,7 +290,7 @@
filter->to = mbfl_no2encoding(to);
/* set the vtbl */
- mbfl_convert_filter_select_vtbl(filter);
+ mbfl_convert_filter_reset_vtbl(filter);
/* construct new filter */
(*filter->filter_ctor)(filter);
@@ -460,7 +459,7 @@
}
-void mbfl_convert_filter_select_vtbl(mbfl_convert_filter *filter)
+static void mbfl_convert_filter_reset_vtbl(mbfl_convert_filter *filter)
{
const struct mbfl_convert_vtbl *vtbl;
1.1.2.2 +0 -1 libmbfl/mbfl/mbfl_convert.h
Index: mbfl_convert.h
===================================================================
RCS file: /cvsroot/php-i18n/libmbfl/mbfl/mbfl_convert.h,v
retrieving revision 1.1.2.1
retrieving revision 1.1.2.2
diff -u -r1.1.2.1 -r1.1.2.2
--- mbfl_convert.h 25 Aug 2003 01:15:33 -0000 1.1.2.1
+++ mbfl_convert.h 4 Feb 2004 05:56:44 -0000 1.1.2.2
@@ -76,7 +76,6 @@
MBFLAPI extern void mbfl_convert_filter_reset(mbfl_convert_filter *filter, enum mbfl_no_encoding from, enum mbfl_no_encoding to);
MBFLAPI extern void mbfl_convert_filter_copy(mbfl_convert_filter *src, mbfl_convert_filter *dist);
MBFLAPI extern int mbfl_filt_conv_illegal_output(int c, mbfl_convert_filter *filter);
-MBFLAPI extern void mbfl_convert_filter_select_vtbl(mbfl_convert_filter *filter);
MBFLAPI extern const struct mbfl_convert_vtbl * mbfl_convert_filter_get_vtbl(enum mbfl_no_encoding from, enum mbfl_no_encoding to);
MBFLAPI extern void mbfl_filt_conv_common_ctor(mbfl_convert_filter *filter);
1.1.2.3 +31 -19 libmbfl/mbfl/mbfl_ident.c
Index: mbfl_ident.c
===================================================================
RCS file: /cvsroot/php-i18n/libmbfl/mbfl/mbfl_ident.c,v
retrieving revision 1.1.2.2
retrieving revision 1.1.2.3
diff -u -r1.1.2.2 -r1.1.2.3
--- mbfl_ident.c 4 Feb 2004 01:58:09 -0000 1.1.2.2
+++ mbfl_ident.c 4 Feb 2004 05:56:44 -0000 1.1.2.3
@@ -154,22 +154,9 @@
return vtbl;
}
-void mbfl_identify_filter_select_vtbl(mbfl_identify_filter *filter)
-{
- const struct mbfl_identify_vtbl *vtbl;
-
- vtbl = mbfl_identify_filter_get_vtbl(filter->encoding->no_encoding);
- if (vtbl == NULL) {
- vtbl = &vtbl_identify_false;
- }
- filter->filter_ctor = vtbl->filter_ctor;
- filter->filter_dtor = vtbl->filter_dtor;
- filter->filter_function = vtbl->filter_function;
-}
-
mbfl_identify_filter *mbfl_identify_filter_new(enum mbfl_no_encoding encoding)
{
- mbfl_identify_filter * filter;
+ mbfl_identify_filter *filter;
/* allocate */
filter = (mbfl_identify_filter *)mbfl_malloc(sizeof(mbfl_identify_filter));
@@ -177,6 +164,18 @@
return NULL;
}
+ if (mbfl_identify_filter_init(filter, encoding)) {
+ mbfl_free(filter);
+ return NULL;
+ }
+
+ return filter;
+}
+
+int mbfl_identify_filter_init(mbfl_identify_filter *filter, enum mbfl_no_encoding encoding)
+{
+ const struct mbfl_identify_vtbl *vtbl;
+
/* encoding structure */
filter->encoding = mbfl_no2encoding(encoding);
if (filter->encoding == NULL) {
@@ -188,20 +187,33 @@
filter->score = 0;
/* setup the function table */
- mbfl_identify_filter_select_vtbl(filter);
+ vtbl = mbfl_identify_filter_get_vtbl(filter->encoding->no_encoding);
+ if (vtbl == NULL) {
+ vtbl = &vtbl_identify_false;
+ }
+ filter->filter_ctor = vtbl->filter_ctor;
+ filter->filter_dtor = vtbl->filter_dtor;
+ filter->filter_function = vtbl->filter_function;
/* constructor */
(*filter->filter_ctor)(filter);
- return filter;
+ return 0;
}
void mbfl_identify_filter_delete(mbfl_identify_filter *filter)
{
- if (filter) {
- (*filter->filter_dtor)(filter);
- mbfl_free((void*)filter);
+ if (filter == NULL) {
+ return;
}
+
+ mbfl_identify_filter_cleanup(filter);
+ mbfl_free((void*)filter);
+}
+
+void mbfl_identify_filter_cleanup(mbfl_identify_filter *filter)
+{
+ (*filter->filter_dtor)(filter);
}
void mbfl_filt_ident_common_ctor(mbfl_identify_filter *filter)
1.1.2.3 +2 -1 libmbfl/mbfl/mbfl_ident.h
Index: mbfl_ident.h
===================================================================
RCS file: /cvsroot/php-i18n/libmbfl/mbfl/mbfl_ident.h,v
retrieving revision 1.1.2.2
retrieving revision 1.1.2.3
diff -u -r1.1.2.2 -r1.1.2.3
--- mbfl_ident.h 4 Feb 2004 01:58:09 -0000 1.1.2.2
+++ mbfl_ident.h 4 Feb 2004 05:56:44 -0000 1.1.2.3
@@ -57,9 +57,10 @@
};
MBFLAPI extern const struct mbfl_identify_vtbl * mbfl_identify_filter_get_vtbl(enum mbfl_no_encoding encoding);
-MBFLAPI extern void mbfl_identify_filter_select_vtbl(mbfl_identify_filter *filter);
MBFLAPI extern mbfl_identify_filter * mbfl_identify_filter_new(enum mbfl_no_encoding encoding);
MBFLAPI extern void mbfl_identify_filter_delete(mbfl_identify_filter *filter);
+MBFLAPI extern mbfl_identify_filter_init(mbfl_identify_filter *filter, enum mbfl_no_encoding encoding);
+MBFLAPI void mbfl_identify_filter_cleanup(mbfl_identify_filter *filter);
MBFLAPI extern void mbfl_filt_ident_common_ctor(mbfl_identify_filter *filter);
MBFLAPI extern void mbfl_filt_ident_common_dtor(mbfl_identify_filter *filter);