X-Git-Url: https://git.distorted.org.uk/~mdw/sgt/charset/blobdiff_plain/032fbecf6f09e462226c96a03965ed12e60e3034..36eb75644933f0cfb984f0509966a4015b93874c:/charset.h

diff --git a/charset.h b/charset.h
index 3b19807..596f05f 100644
--- 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,
@@ -134,6 +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,
@@ -164,6 +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,
@@ -197,6 +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