I've just had some spam in Windows-874, a Thai SBCS. Add libcharset

[sgt/charset] / charset.h

diff --git a/charset.h b/charset.h
index 0dca159..596f05f 100644 (file)
--- a/charset.h
+++ b/charset.h
@@ -33,6 +33,8 @@ typedef enum {
     CS_ISO8859_16,
     CS_CP437,
     CS_CP850,
+    CS_CP866,
+    CS_CP874,
     CS_CP1250,
     CS_CP1251,
     CS_CP1252,
@@ -45,6 +47,7 @@ typedef enum {
     CS_KOI8_R,
     CS_KOI8_U,
     CS_KOI8_RU,
+    CS_JISX0201,
     CS_MAC_ROMAN,
     CS_MAC_TURKISH,
     CS_MAC_CROATIAN,
@@ -83,6 +86,13 @@ typedef enum {
     CS_SHIFT_JIS,
     CS_HZ,
     CS_CP949,
+    CS_PDF,
+    CS_PSSTD,
+    CS_CTEXT,
+    CS_ISO2022,
+    CS_BS4730,
+    CS_DEC_GRAPHICS,
+    CS_EUC_TW
 } charset_t;
 
 typedef struct {
@@ -126,6 +136,13 @@ extern const charset_state charset_init_state;
  * NULL, `errlen' will be ignored, and the library will choose
  * something sensible to do on its own. For Unicode, this will be
  * U+FFFD (REPLACEMENT CHARACTER).
+ * 
+ * `output' may be NULL, in which case the entire translation will
+ * be performed in theory (e.g. a dry run to work out how much
+ * space needs to be allocated for the real thing). `outlen' may
+ * also be negative, indicating an unlimited buffer length
+ * (although this is almost certainly unwise if `output' is _not_
+ * NULL).
  */
 
 int charset_to_unicode(const char **input, int *inlen,
@@ -156,6 +173,13 @@ int charset_to_unicode(const char **input, int *inlen,
  * If `input' is NULL, this routine will output the necessary bytes
  * to reset the encoding state in any way which might be required
  * at the end of an output piece of text.
+ * 
+ * `output' may be NULL, in which case the entire translation will
+ * be performed in theory (e.g. a dry run to work out how much
+ * space needs to be allocated for the real thing). `outlen' may
+ * also be negative, indicating an unlimited buffer length
+ * (although this is almost certainly unwise if `output' is _not_
+ * NULL).
  */
 
 int charset_from_unicode(const wchar_t **input, int *inlen,
@@ -189,6 +213,13 @@ int charset_from_macenc(int script, int region, int sysvers,
                        const char *fontname);
 
 /*
+ * Convert GNU Emacs coding system symbol to and from our charset
+ * identifiers.
+ */
+const char *charset_to_emacsenc(int charset);
+int charset_from_emacsenc(const char *name);
+
+/*
  * Upgrade a charset identifier to a superset charset which is
  * often confused with it. For example, people whose MUAs report
  * their mail as ASCII or ISO8859-1 often in practice turn out to
@@ -206,4 +237,15 @@ int charset_upgrade(int charset);
  */
 int charset_contains_ascii(int charset);
 
+/*
+ * This function tries to deduce the CS_* identifier of the charset
+ * used in the current C locale. It falls back to CS_ASCII if it
+ * can't figure it out at all, so it will always return a valid
+ * charset.
+ * 
+ * (Note that you should have already called setlocale(LC_CTYPE,
+ * "") to guarantee that this function will do the right thing.)
+ */
+int charset_from_locale(void);
+
 #endif /* charset_charset_h */