Silly of me to overlook it: another obvious way you might like to

[sgt/charset] / charset.h

diff --git a/charset.h b/charset.h
index ebaafc9..bbc69e4 100644 (file)
--- a/charset.h
+++ b/charset.h
@@ -33,6 +33,9 @@ typedef enum {
     CS_ISO8859_16,
     CS_CP437,
     CS_CP850,
+    CS_CP852,
+    CS_CP866,
+    CS_CP874,
     CS_CP1250,
     CS_CP1251,
     CS_CP1252,
@@ -85,7 +88,13 @@ typedef enum {
     CS_HZ,
     CS_CP949,
     CS_PDF,
+    CS_PSSTD,
     CS_CTEXT,
+    CS_ISO2022,
+    CS_BS4730,
+    CS_DEC_GRAPHICS,
+    CS_EUC_TW,
+    CS_LIMIT               /* dummy value indicating extent of enum */
 } charset_t;
 
 typedef struct {
@@ -129,6 +138,13 @@ 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,
@@ -159,6 +175,13 @@ 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,
@@ -192,6 +215,13 @@ 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
@@ -220,4 +250,19 @@ int charset_contains_ascii(int charset);
  */
 int charset_from_locale(void);
 
+/*
+ * This function simply reports whether a charset identifier
+ * corresponds to an actually usable charset. Not everything in the
+ * above enum does: CS_NONE, for a start, and occasionally other slots
+ * in the enum are reserved before they actually go into service.
+ *
+ * This function permits clients to iterate over _all_ supported
+ * charsets by means of a loop such as
+ *
+ *     for (cs = 0; cs < CS_LIMIT; cs++)
+ *         if (charset_exists(cs))
+ *             do_stuff_with(cs);
+ */
+int charset_exists(int charset);
+
 #endif /* charset_charset_h */