string-array: New functions for comparing string arrays.

[pspp] / src / libpspp / i18n.c

diff --git a/src/libpspp/i18n.c b/src/libpspp/i18n.c
index 69162f14f07c72b35e55537f2bd0623f2bb20d5c..3faadcbb87c770395f5ce0c86fdef509df50c612 100644 (file)
--- a/src/libpspp/i18n.c
+++ b/src/libpspp/i18n.c
@@ -29,11 +29,13 @@
 #include <string.h>
 #include <unicase.h>
 #include <unigbrk.h>
+#include <uniwidth.h>
 
 #include "libpspp/assertion.h"
 #include "libpspp/compiler.h"
 #include "libpspp/hmapx.h"
 #include "libpspp/hash-functions.h"
+#include "libpspp/misc.h"
 #include "libpspp/pool.h"
 #include "libpspp/str.h"
 #include "libpspp/version.h"
@@ -501,6 +503,55 @@ utf8_encoding_concat_len (const char *head, const char *tail,
   return prefix_len + tail_len;
 }
 
+/* Returns the number of display columns that would be occupied by the LENGTH
+   bytes of UTF-8 starting at S. */
+size_t
+utf8_count_columns (const char *s_, size_t length)
+{
+  const uint8_t *s = CHAR_CAST (const uint8_t *, s_);
+
+  size_t columns = 0;
+  for (int ofs = 0; ofs < length; )
+    {
+      ucs4_t uc;
+      ofs += u8_mbtouc (&uc, s + ofs, length - ofs);
+      if (uc != '\t')
+        {
+          int width = uc_width (uc, "UTF-8");
+          if (width > 0)
+            columns += width;
+        }
+      else
+        columns = ROUND_UP (columns + 1, 8);
+    }
+  return columns;
+}
+
+/* Returns the byte offset in LENGTH-byte UTF-8 string S that is N_COLUMNS
+   display columns into the string. */
+size_t
+utf8_columns_to_bytes (const char *s_, size_t length, size_t n_columns)
+{
+  const uint8_t *s = CHAR_CAST (const uint8_t *, s_);
+
+  size_t columns = 0;
+  int ofs;
+  for (ofs = 0; ofs < length && columns < n_columns; )
+    {
+      ucs4_t uc;
+      ofs += u8_mbtouc (&uc, s + ofs, length - ofs);
+      if (uc != '\t')
+        {
+          int width = uc_width (uc, "UTF-8");
+          if (width > 0)
+            columns += width;
+        }
+      else
+        columns = ROUND_UP (columns + 1, 8);
+    }
+  return ofs;
+}
+
 /* Returns an allocated, null-terminated string, owned by the caller,
    containing as many characters[*] from the beginning of S that would fit
    within MAX_LEN bytes if the returned string were to be re-encoded in
@@ -829,7 +880,15 @@ utf8_hash_case_bytes (const char *s, size_t n, unsigned int basis)
 unsigned int
 utf8_hash_case_string (const char *s, unsigned int basis)
 {
-  return utf8_hash_case_bytes (s, strlen (s), basis);
+  return utf8_hash_case_substring (ss_cstr (s), basis);
+}
+
+/* Returns a hash value for UTF-8 string S, with lowercase and uppercase
+   letters treated as equal, starting from BASIS. */
+unsigned int
+utf8_hash_case_substring (struct substring s, unsigned int basis)
+{
+  return utf8_hash_case_bytes (s.string, s.length, basis);
 }
 
 /* Compares UTF-8 strings A and B case-insensitively.
@@ -837,7 +896,13 @@ utf8_hash_case_string (const char *s, unsigned int basis)
 int
 utf8_strcasecmp (const char *a, const char *b)
 {
-  return utf8_strncasecmp (a, strlen (a), b, strlen (b));
+  return utf8_sscasecmp (ss_cstr (a), ss_cstr (b));
+}
+
+int
+utf8_sscasecmp (struct substring a, struct substring b)
+{
+  return utf8_strncasecmp (a.string, a.length, b.string, b.length);
 }
 
 /* Compares UTF-8 strings A (with length AN) and B (with length BN)