encoding-guesser: Fall back to windows-1252 when UTF-8 can't be right.

[pspp] / src / libpspp / str.c

diff --git a/src/libpspp/str.c b/src/libpspp/str.c
index ba4a26f1426ac51552dd31885d5cdfff5fea321d..79e9ea1e14dae1def667b3f4edd1ea60f4e0ac5d 100644 (file)
--- a/src/libpspp/str.c
+++ b/src/libpspp/str.c
@@ -1,5 +1,5 @@
 /* PSPP - a program for statistical analysis.
-   Copyright (C) 1997-9, 2000, 2006, 2009, 2010, 2011 Free Software Foundation, Inc.
+   Copyright (C) 1997-9, 2000, 2006, 2009, 2010, 2011, 2012 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -270,17 +270,22 @@ str_format_26adic (unsigned long int number, char buffer[], size_t size)
   while (number-- > 0)
     {
       if (length >= size)
-        return false;
+        goto overflow;
       buffer[length++] = "ABCDEFGHIJKLMNOPQRSTUVWXYZ"[number % 26];
       number /= 26;
     }
 
   if (length >= size)
-    return false;
+    goto overflow;
   buffer[length] = '\0';
 
   buf_reverse (buffer, length);
   return true;
+
+overflow:
+  if (length > 0)
+    buffer[0] = '\0';
+  return false;
 }
 
 /* Sets the SIZE bytes starting at BLOCK to C,
@@ -347,15 +352,16 @@ ss_realloc (struct substring *ss, size_t size)
   ss->string = xrealloc (ss->string, size);
 }
 
-/* Makes a pool_alloc_unaligned()'d copy of the contents of OLD
-   in POOL, and stores it in NEW. */
+/* Makes a pool_alloc_unaligned()'d, null-terminated copy of the contents of
+   OLD in POOL, and stores it in NEW. */
 void
 ss_alloc_substring_pool (struct substring *new, struct substring old,
                          struct pool *pool)
 {
-  new->string = pool_alloc_unaligned (pool, old.length);
+  new->string = pool_alloc_unaligned (pool, old.length + 1);
   new->length = old.length;
   memcpy (new->string, old.string, old.length);
+  new->string[old.length] = '\0';
 }
 
 /* Allocates room for a CNT-byte string in NEW in POOL. */
@@ -485,11 +491,15 @@ bool
 ss_tokenize (struct substring ss, struct substring delimiters,
              size_t *save_idx, struct substring *token)
 {
+  bool found_token;
+
   ss_advance (&ss, *save_idx);
   *save_idx += ss_ltrim (&ss, delimiters);
   ss_get_bytes (&ss, ss_cspan (ss, delimiters), token);
-  *save_idx += ss_length (*token) + 1;
-  return ss_length (*token) > 0;
+
+  found_token = ss_length (*token) > 0;
+  *save_idx += ss_length (*token) + found_token;
+  return found_token;
 }
 
 /* Removes the first CNT bytes from SS. */
@@ -1454,6 +1464,30 @@ ds_put_uninit (struct string *st, size_t incr)
   return end;
 }
 
+/* Moves the bytes in ST following offset OFS + OLD_LEN in ST to offset OFS +
+   NEW_LEN and returns the byte at offset OFS.  The first min(OLD_LEN, NEW_LEN)
+   bytes at the returned position are unchanged; if NEW_LEN > OLD_LEN then the
+   following NEW_LEN - OLD_LEN bytes are initially indeterminate.
+
+   The intention is that the caller should write NEW_LEN bytes at the returned
+   position, to effectively replace the OLD_LEN bytes previously at that
+   position. */
+char *
+ds_splice_uninit (struct string *st,
+                  size_t ofs, size_t old_len, size_t new_len)
+{
+  if (new_len != old_len)
+    {
+      if (new_len > old_len)
+        ds_extend (st, ds_length (st) + (new_len - old_len));
+      memmove (ds_data (st) + (ofs + new_len),
+               ds_data (st) + (ofs + old_len),
+               ds_length (st) - (ofs + old_len));
+      st->ss.length += new_len - old_len;
+    }
+  return ds_data (st) + ofs;
+}
+
 /* Formats FORMAT as a printf string and appends the result to ST. */
 void
 ds_put_format (struct string *st, const char *format, ...)