tutf8e: Refresh from upstream

Signed-off-by: Nigel Stewart <[email protected]>

Author: nigels-com

lib/tutf8e/codegen.py

@@ -6,118 +6,68 @@
 #include <stdint.h>  /* uint16_t */
 
 /* Internal API */
+
 extern int tutf8e_string_length(const uint16_t *table, const char *i, size_t *ilen, size_t *olen);
-extern int tutf8e_buffer_length(const uint16_t *table, const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_encode(const uint16_t *table, const char *i, size_t ilen, char *output, size_t *olen);
+extern int tutf8e_string_encode(const uint16_t *table, const char *i, char *o, size_t *olen);
+
+extern int tutf8e_buffer_length(const uint16_t *table, const char *i, size_t ilen, size_t *olen);
+extern int tutf8e_buffer_encode(const uint16_t *table, const char *i, size_t ilen, char *o, size_t *olen);
+
+/* Generic API */
 
-/* External API */
+typedef void *TUTF8encoder;
+
+extern TUTF8encoder tutf8e_encoder(const char *encoding);
 
 #define TUTF8E_OK      0 /* Success                    */
 #define TUTF8E_INVALID 1 /* Invalid input character    */
 #define TUTF8E_TOOLONG 2 /* Insufficient output buffer */
 
-/* Encode NUL-terminated string to UTF8 */
-extern int tutf8e_string_encode_iso_8859_1  (char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_iso_8859_10 (char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_iso_8859_11 (char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_iso_8859_13 (char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_iso_8859_14 (char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_iso_8859_15 (char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_iso_8859_16 (char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_iso_8859_2  (char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_iso_8859_3  (char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_iso_8859_4  (char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_iso_8859_5  (char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_iso_8859_6  (char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_iso_8859_7  (char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_iso_8859_8  (char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_iso_8859_9  (char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_windows_1250(char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_windows_1251(char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_windows_1252(char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_windows_1253(char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_windows_1254(char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_windows_1255(char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_windows_1256(char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_windows_1257(char *output, size_t olen, const char *input);
-extern int tutf8e_string_encode_windows_1258(char *output, size_t olen, const char *input);
-
-/* Encode NUL-terminated string to UTF8, realloc as necessary */
-extern char * tutf8e_string_encode_iso_8859_1_realloc(char *input);
-extern char * tutf8e_string_encode_iso_8859_10_realloc(char *input);
-extern char * tutf8e_string_encode_iso_8859_11_realloc(char *input);
-extern char * tutf8e_string_encode_iso_8859_13_realloc(char *input);
-extern char * tutf8e_string_encode_iso_8859_14_realloc(char *input);
-extern char * tutf8e_string_encode_iso_8859_15_realloc(char *input);
-extern char * tutf8e_string_encode_iso_8859_16_realloc(char *input);
-extern char * tutf8e_string_encode_iso_8859_2_realloc(char *input);
-extern char * tutf8e_string_encode_iso_8859_3_realloc(char *input);
-extern char * tutf8e_string_encode_iso_8859_4_realloc(char *input);
-extern char * tutf8e_string_encode_iso_8859_5_realloc(char *input);
-extern char * tutf8e_string_encode_iso_8859_6_realloc(char *input);
-extern char * tutf8e_string_encode_iso_8859_7_realloc(char *input);
-extern char * tutf8e_string_encode_iso_8859_8_realloc(char *input);
-extern char * tutf8e_string_encode_iso_8859_9_realloc(char *input);
-extern char * tutf8e_string_encode_windows_1250_realloc(char *input);
-extern char * tutf8e_string_encode_windows_1251_realloc(char *input);
-extern char * tutf8e_string_encode_windows_1252_realloc(char *input);
-extern char * tutf8e_string_encode_windows_1253_realloc(char *input);
-extern char * tutf8e_string_encode_windows_1254_realloc(char *input);
-extern char * tutf8e_string_encode_windows_1255_realloc(char *input);
-extern char * tutf8e_string_encode_windows_1256_realloc(char *input);
-extern char * tutf8e_string_encode_windows_1257_realloc(char *input);
-extern char * tutf8e_string_encode_windows_1258_realloc(char *input);
-
-/* Buffer length as UTF8 */
-extern int tutf8e_buffer_length_iso_8859_1  (const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_iso_8859_10 (const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_iso_8859_11 (const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_iso_8859_13 (const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_iso_8859_14 (const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_iso_8859_15 (const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_iso_8859_16 (const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_iso_8859_2  (const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_iso_8859_3  (const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_iso_8859_4  (const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_iso_8859_5  (const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_iso_8859_6  (const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_iso_8859_7  (const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_iso_8859_8  (const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_iso_8859_9  (const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_windows_1250(const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_windows_1251(const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_windows_1252(const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_windows_1253(const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_windows_1254(const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_windows_1255(const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_windows_1256(const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_windows_1257(const char *i, size_t ilen, size_t *length);
-extern int tutf8e_buffer_length_windows_1258(const char *i, size_t ilen, size_t *length);
-
-/* Encode buffer to UTF8 */
-extern int tutf8e_buffer_encode_iso_8859_1  (char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_iso_8859_10 (char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_iso_8859_11 (char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_iso_8859_13 (char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_iso_8859_14 (char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_iso_8859_15 (char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_iso_8859_16 (char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_iso_8859_2  (char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_iso_8859_3  (char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_iso_8859_4  (char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_iso_8859_5  (char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_iso_8859_6  (char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_iso_8859_7  (char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_iso_8859_8  (char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_iso_8859_9  (char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_windows_1250(char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_windows_1251(char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_windows_1252(char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_windows_1253(char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_windows_1254(char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_windows_1255(char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_windows_1256(char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_windows_1257(char *output, size_t *olen, const char *input, size_t ilen);
-extern int tutf8e_buffer_encode_windows_1258(char *output, size_t *olen, const char *input, size_t ilen);
+static inline int tutf8e_encoder_string_length(const TUTF8encoder encoder, const char *i, size_t *ilen, size_t *olen)
+{
+  return tutf8e_string_length((const uint16_t *) encoder, i, ilen, olen);
+}
+
+static inline int tutf8e_encoder_string_encode(const TUTF8encoder encoder, const char *i, char *o, size_t *olen)
+{
+  return tutf8e_string_encode((const uint16_t *) encoder, i, o, olen);
+}
+
+static inline int tutf8e_encoder_buffer_length(const TUTF8encoder encoder, const char *i, size_t ilen, size_t *length)
+{
+  return tutf8e_buffer_length((const uint16_t *) encoder, i, ilen, length);
+}
+
+static inline int tutf8e_encoder_buffer_encode(const TUTF8encoder encoder, const char *i, size_t ilen, char *o, size_t *olen)
+{
+  return tutf8e_buffer_encode((const uint16_t *) encoder, i, ilen, o, olen);
+}
+
+/* Supported encoders */
+
+extern const TUTF8encoder tutf8e_encoder_iso_8859_1;
+extern const TUTF8encoder tutf8e_encoder_iso_8859_10;
+extern const TUTF8encoder tutf8e_encoder_iso_8859_11;
+extern const TUTF8encoder tutf8e_encoder_iso_8859_13;
+extern const TUTF8encoder tutf8e_encoder_iso_8859_14;
+extern const TUTF8encoder tutf8e_encoder_iso_8859_15;
+extern const TUTF8encoder tutf8e_encoder_iso_8859_16;
+extern const TUTF8encoder tutf8e_encoder_iso_8859_2;
+extern const TUTF8encoder tutf8e_encoder_iso_8859_3;
+extern const TUTF8encoder tutf8e_encoder_iso_8859_4;
+extern const TUTF8encoder tutf8e_encoder_iso_8859_5;
+extern const TUTF8encoder tutf8e_encoder_iso_8859_6;
+extern const TUTF8encoder tutf8e_encoder_iso_8859_7;
+extern const TUTF8encoder tutf8e_encoder_iso_8859_8;
+extern const TUTF8encoder tutf8e_encoder_iso_8859_9;
+extern const TUTF8encoder tutf8e_encoder_windows_1250;
+extern const TUTF8encoder tutf8e_encoder_windows_1251;
+extern const TUTF8encoder tutf8e_encoder_windows_1252;
+extern const TUTF8encoder tutf8e_encoder_windows_1253;
+extern const TUTF8encoder tutf8e_encoder_windows_1254;
+extern const TUTF8encoder tutf8e_encoder_windows_1255;
+extern const TUTF8encoder tutf8e_encoder_windows_1256;
+extern const TUTF8encoder tutf8e_encoder_windows_1257;
+extern const TUTF8encoder tutf8e_encoder_windows_1258;
 
 #endif

lib/tutf8e/include/tutf8e.h

@@ -1,9 +1,6 @@
 #include <tutf8e.h>
 
-#include <string.h> /* strlen */
-#include <stdlib.h> /* malloc/free */
-
-static const uint16_t iso_8859_1_utf8[256] =
+static const uint16_t tutf8e_iso_8859_1_utf8[256] =
 {
   0x0000, 0x0001, 0x0002, 0x0003, 0x0004, 0x0005, 0x0006, 0x0007, 0x0008, 0x0009, 0x000a, 0x000b, 0x000c, 0x000d, 0x000e, 0x000f,
   0x0010, 0x0011, 0x0012, 0x0013, 0x0014, 0x0015, 0x0016, 0x0017, 0x0018, 0x0019, 0x001a, 0x001b, 0x001c, 0x001d, 0x001e, 0x001f,
@@ -23,34 +20,4 @@ static const uint16_t iso_8859_1_utf8[256] =
   0x00f0, 0x00f1, 0x00f2, 0x00f3, 0x00f4, 0x00f5, 0x00f6, 0x00f7, 0x00f8, 0x00f9, 0x00fa, 0x00fb, 0x00fc, 0x00fd, 0x00fe, 0x00ff,
 };
 
-int tutf8e_string_encode_iso_8859_1(char *output, size_t olen, const char *input)
-{
-  size_t len = strlen(input) + 1;
-  return tutf8e_buffer_encode(iso_8859_1_utf8, input, len, output, &olen);
-}
-
-int tutf8e_buffer_length_iso_8859_1(const char *i, size_t ilen, size_t *length)
-{
-  return tutf8e_buffer_length(iso_8859_1_utf8, i, ilen, length);
-}
-
-int tutf8e_buffer_encode_iso_8859_1(char *output, size_t *olen, const char *input, size_t ilen)
-{
-  return tutf8e_buffer_encode(iso_8859_1_utf8, input, ilen, output, olen);
-}
-
-char * tutf8e_string_encode_iso_8859_1_realloc(char *input)
-{
-  size_t ilen = 0;
-  size_t olen = 0;
-  if (input && !tutf8e_string_length(iso_8859_1_utf8, input, &ilen, &olen) && ilen && olen && ilen!=olen) {
-    char * output = malloc(olen + 1);
-    if (output && !tutf8e_buffer_encode(iso_8859_1_utf8, input, ilen, output, &olen)) {
-      free(input);
-      output[olen] = 0;
-      return output;
-    }
-    free(output);
-  }
-  return input;
-}
+const TUTF8encoder tutf8e_encoder_iso_8859_1 = (TUTF8encoder) tutf8e_iso_8859_1_utf8;

lib/tutf8e/src/iso_8859_1.c

@@ -1,9 +1,6 @@
 #include <tutf8e.h>
 
-#include <string.h> /* strlen */
-#include <stdlib.h> /* malloc/free */
-
-static const uint16_t iso_8859_10_utf8[256] =
+static const uint16_t tutf8e_iso_8859_10_utf8[256] =
 {
   0x0000, 0x0001, 0x0002, 0x0003, 0x0004, 0x0005, 0x0006, 0x0007, 0x0008, 0x0009, 0x000a, 0x000b, 0x000c, 0x000d, 0x000e, 0x000f,
   0x0010, 0x0011, 0x0012, 0x0013, 0x0014, 0x0015, 0x0016, 0x0017, 0x0018, 0x0019, 0x001a, 0x001b, 0x001c, 0x001d, 0x001e, 0x001f,
@@ -23,34 +20,4 @@ static const uint16_t iso_8859_10_utf8[256] =
   0x00f0, 0x0146, 0x014d, 0x00f3, 0x00f4, 0x00f5, 0x00f6, 0x0169, 0x00f8, 0x0173, 0x00fa, 0x00fb, 0x00fc, 0x00fd, 0x00fe, 0x0138,
 };
 
-int tutf8e_string_encode_iso_8859_10(char *output, size_t olen, const char *input)
-{
-  size_t len = strlen(input) + 1;
-  return tutf8e_buffer_encode(iso_8859_10_utf8, input, len, output, &olen);
-}
-
-int tutf8e_buffer_length_iso_8859_10(const char *i, size_t ilen, size_t *length)
-{
-  return tutf8e_buffer_length(iso_8859_10_utf8, i, ilen, length);
-}
-
-int tutf8e_buffer_encode_iso_8859_10(char *output, size_t *olen, const char *input, size_t ilen)
-{
-  return tutf8e_buffer_encode(iso_8859_10_utf8, input, ilen, output, olen);
-}
-
-char * tutf8e_string_encode_iso_8859_10_realloc(char *input)
-{
-  size_t ilen = 0;
-  size_t olen = 0;
-  if (input && !tutf8e_string_length(iso_8859_10_utf8, input, &ilen, &olen) && ilen && olen && ilen!=olen) {
-    char * output = malloc(olen + 1);
-    if (output && !tutf8e_buffer_encode(iso_8859_10_utf8, input, ilen, output, &olen)) {
-      free(input);
-      output[olen] = 0;
-      return output;
-    }
-    free(output);
-  }
-  return input;
-}
+const TUTF8encoder tutf8e_encoder_iso_8859_10 = (TUTF8encoder) tutf8e_iso_8859_10_utf8;

lib/tutf8e/src/iso_8859_10.c

@@ -1,9 +1,6 @@
 #include <tutf8e.h>
 
-#include <string.h> /* strlen */
-#include <stdlib.h> /* malloc/free */
-
-static const uint16_t iso_8859_11_utf8[256] =
+static const uint16_t tutf8e_iso_8859_11_utf8[256] =
 {
   0x0000, 0x0001, 0x0002, 0x0003, 0x0004, 0x0005, 0x0006, 0x0007, 0x0008, 0x0009, 0x000a, 0x000b, 0x000c, 0x000d, 0x000e, 0x000f,
   0x0010, 0x0011, 0x0012, 0x0013, 0x0014, 0x0015, 0x0016, 0x0017, 0x0018, 0x0019, 0x001a, 0x001b, 0x001c, 0x001d, 0x001e, 0x001f,
@@ -23,34 +20,4 @@ static const uint16_t iso_8859_11_utf8[256] =
   0x0e50, 0x0e51, 0x0e52, 0x0e53, 0x0e54, 0x0e55, 0x0e56, 0x0e57, 0x0e58, 0x0e59, 0x0e5a, 0x0e5b, 0xffff, 0xffff, 0xffff, 0xffff,
 };
 
-int tutf8e_string_encode_iso_8859_11(char *output, size_t olen, const char *input)
-{
-  size_t len = strlen(input) + 1;
-  return tutf8e_buffer_encode(iso_8859_11_utf8, input, len, output, &olen);
-}
-
-int tutf8e_buffer_length_iso_8859_11(const char *i, size_t ilen, size_t *length)
-{
-  return tutf8e_buffer_length(iso_8859_11_utf8, i, ilen, length);
-}
-
-int tutf8e_buffer_encode_iso_8859_11(char *output, size_t *olen, const char *input, size_t ilen)
-{
-  return tutf8e_buffer_encode(iso_8859_11_utf8, input, ilen, output, olen);
-}
-
-char * tutf8e_string_encode_iso_8859_11_realloc(char *input)
-{
-  size_t ilen = 0;
-  size_t olen = 0;
-  if (input && !tutf8e_string_length(iso_8859_11_utf8, input, &ilen, &olen) && ilen && olen && ilen!=olen) {
-    char * output = malloc(olen + 1);
-    if (output && !tutf8e_buffer_encode(iso_8859_11_utf8, input, ilen, output, &olen)) {
-      free(input);
-      output[olen] = 0;
-      return output;
-    }
-    free(output);
-  }
-  return input;
-}
+const TUTF8encoder tutf8e_encoder_iso_8859_11 = (TUTF8encoder) tutf8e_iso_8859_11_utf8;

lib/tutf8e/src/iso_8859_11.c

@@ -1,9 +1,6 @@
 #include <tutf8e.h>
 
-#include <string.h> /* strlen */
-#include <stdlib.h> /* malloc/free */
-
-static const uint16_t iso_8859_13_utf8[256] =
+static const uint16_t tutf8e_iso_8859_13_utf8[256] =
 {
   0x0000, 0x0001, 0x0002, 0x0003, 0x0004, 0x0005, 0x0006, 0x0007, 0x0008, 0x0009, 0x000a, 0x000b, 0x000c, 0x000d, 0x000e, 0x000f,
   0x0010, 0x0011, 0x0012, 0x0013, 0x0014, 0x0015, 0x0016, 0x0017, 0x0018, 0x0019, 0x001a, 0x001b, 0x001c, 0x001d, 0x001e, 0x001f,
@@ -23,34 +20,4 @@ static const uint16_t iso_8859_13_utf8[256] =
   0x0161, 0x0144, 0x0146, 0x00f3, 0x014d, 0x00f5, 0x00f6, 0x00f7, 0x0173, 0x0142, 0x015b, 0x016b, 0x00fc, 0x017c, 0x017e, 0x2019,
 };
 
-int tutf8e_string_encode_iso_8859_13(char *output, size_t olen, const char *input)
-{
-  size_t len = strlen(input) + 1;
-  return tutf8e_buffer_encode(iso_8859_13_utf8, input, len, output, &olen);
-}
-
-int tutf8e_buffer_length_iso_8859_13(const char *i, size_t ilen, size_t *length)
-{
-  return tutf8e_buffer_length(iso_8859_13_utf8, i, ilen, length);
-}
-
-int tutf8e_buffer_encode_iso_8859_13(char *output, size_t *olen, const char *input, size_t ilen)
-{
-  return tutf8e_buffer_encode(iso_8859_13_utf8, input, ilen, output, olen);
-}
-
-char * tutf8e_string_encode_iso_8859_13_realloc(char *input)
-{
-  size_t ilen = 0;
-  size_t olen = 0;
-  if (input && !tutf8e_string_length(iso_8859_13_utf8, input, &ilen, &olen) && ilen && olen && ilen!=olen) {
-    char * output = malloc(olen + 1);
-    if (output && !tutf8e_buffer_encode(iso_8859_13_utf8, input, ilen, output, &olen)) {
-      free(input);
-      output[olen] = 0;
-      return output;
-    }
-    free(output);
-  }
-  return input;
-}
+const TUTF8encoder tutf8e_encoder_iso_8859_13 = (TUTF8encoder) tutf8e_iso_8859_13_utf8;