X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Flibpspp%2Fencoding-guesser.c;h=7d10015e2ea2961febf0efb8edb47d66e30033a8;hb=refs%2Fbuilds%2F20110716030503%2Fpspp;hp=9042e93a2fd41db02012d19f206c378748adc7c8;hpb=f3668539947d5baed813a4f8436d6cf36abeedd2;p=pspp

diff --git a/src/libpspp/encoding-guesser.c b/src/libpspp/encoding-guesser.c
index 9042e93a2f..7d10015e2e 100644
--- a/src/libpspp/encoding-guesser.c
+++ b/src/libpspp/encoding-guesser.c
@@ -250,10 +250,6 @@ encoding_guess_head_encoding (const char *encoding,
       || !encoding_guess_tail_is_utf8 (data, n))
     return fallback_encoding;
 
-  if (!c_strcasecmp (fallback_encoding, "UTF-8")
-      || !c_strcasecmp (fallback_encoding, "UTF8"))
-    return "UTF-8";
-
   return "ASCII";
 }
 
@@ -287,3 +283,30 @@ encoding_guess_tail_is_utf8 (const void *data, size_t n)
           : is_all_utf8_text (data, n));
 }
 
+/* Attempts to guess the encoding of a text file based on ENCODING, an encoding
+   name in one of the forms described at the top of encoding-guesser.h, and the
+   SIZE byts in DATA, which contains the entire contents of the file.  Returns
+   the guessed encoding, which might be ENCODING itself or a suffix of it or a
+   statically allocated string.
+
+   Encoding autodetection only takes place if ENCODING actually specifies
+   autodetection.  See encoding-guesser.h for details. */
+const char *
+encoding_guess_whole_file (const char *encoding, const void *text, size_t size)
+{
+  const char *guess;
+
+  guess = encoding_guess_head_encoding (encoding, text, size);
+  if (!strcmp (guess, "ASCII") && encoding_guess_encoding_is_auto (encoding))
+    {
+      size_t ofs = encoding_guess_count_ascii (text, size);
+      if (ofs < size)
+        return encoding_guess_tail_encoding (encoding,
+                                             (const char *) text + ofs,
+                                             size - ofs);
+      else
+        return encoding_guess_parse_encoding (encoding);
+    }
+  else
+    return guess;
+}