hash-functions: New function hash_case_bytes().

This is useful for hashing an arbitrary byte sequence case-insensitively.
Obviously most uses would be better off working with Unicode but we aren't
there yet.

diff --git a/src/libpspp/hash-functions.c b/src/libpspp/hash-functions.c
index 96ff282c85d7fc9684672cd840a5ea1cca33afe8..c43017313a789666ca83bb8373351bc59972a3b0 100644 (file)
--- a/src/libpspp/hash-functions.c
+++ b/src/libpspp/hash-functions.c
@@ -102,13 +102,12 @@ hash_string (const char *s, unsigned int basis)
   return hash_bytes (s, strlen (s), basis);
 }
 
-/* Returns a hash value for null-terminated string S, with
-   lowercase and uppercase letters treated as equal, starting
-   from BASIS. */
+/* Returns a hash value for the N bytes at S, with lowercase and uppercase
+   letters treated as equal, starting from BASIS. */
 unsigned int
-hash_case_string (const char *s, unsigned int basis)
+hash_case_bytes (const void *s_, size_t n, unsigned int basis)
 {
-  size_t n = strlen (s);
+  const char *s = s_;
   uint32_t a, b, c;
   uint32_t tmp[3];
   int i;
@@ -141,6 +140,15 @@ hash_case_string (const char *s, unsigned int basis)
   return c;
 }
 
+/* Returns a hash value for null-terminated string S, with
+   lowercase and uppercase letters treated as equal, starting
+   from BASIS. */
+unsigned int
+hash_case_string (const char *s, unsigned int basis)
+{
+  return hash_case_bytes (s, strlen (s), basis);
+}
+
 /* Returns a hash value for integer X, starting from BASIS. */
 unsigned int
 hash_int (int x, unsigned int basis)

diff --git a/src/libpspp/hash-functions.h b/src/libpspp/hash-functions.h
index bdf3da31ab7578fa464b693fe059b91382957ee3..86414fb1c22dc870e6b0a5ac53ba2f1b3a7805d1 100644 (file)
--- a/src/libpspp/hash-functions.h
+++ b/src/libpspp/hash-functions.h
@@ -21,6 +21,7 @@
 
 unsigned int hash_bytes (const void *, size_t, unsigned int basis);
 unsigned int hash_string (const char *, unsigned int basis);
+unsigned int hash_case_bytes (const void *, size_t, unsigned int basis);
 unsigned int hash_case_string (const char *, unsigned int basis);
 unsigned int hash_int (int, unsigned int basis);
 unsigned int hash_double (double, unsigned int basis);