utf8proc

commit a39c1a6ea287e10c72c8b5d3013d232b7f85af3c
parent 5fbd15b4ce9db9b12b48cf3b94c70280bf20a13b
Author: Steven G. Johnson <stevenj@alum.mit.edu>
Date:   Sat, 19 Jul 2014 14:55:25 -0400

added GNU libunistring benchmark

Diffstat:
M
.gitignore
 | 
7
+++++--
M
bench/Makefile
 | 
8
+++++++-
A
bench/unistring.c
 | 
60
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

3 files changed, 72 insertions(+), 3 deletions(-)
diff --git a/.gitignore b/.gitignore
@@ -10,5 +10,8 @@
 *.dSYM
 *.txt
 *.out
-bench
-icu
+bench/bench
+bench/icu
+bench/unistring
+
+
diff --git a/bench/Makefile b/bench/Makefile
@@ -26,8 +26,14 @@ icu: icu.o util.o
 icu.out: $(DATAFILES) icu
 	./icu $(DATAFILES) > $@
 
+unistring: unistring.o util.o
+	$(CC) $(CFLAGS) $(LDFLAGS) -o $@ unistring.o util.o -lunistring
+
+unistring.out: $(DATAFILES) unistring
+	./unistring $(DATAFILES) > $@
+
 .c.o:
 	$(CC) $(CPPFLAGS) -I.. $(CFLAGS) -c -o $@ $<
 
 clean:
-	rm -rf *.o *.txt bench *.out icu
+	rm -rf *.o *.txt bench *.out icu unistring
diff --git a/bench/unistring.c b/bench/unistring.c
@@ -0,0 +1,60 @@
+/* comparitive benchmark of GNU libunistring */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+/* libunistring */
+#include <unistr.h>
+#include <uninorm.h>
+
+#include "util.h"
+
+int main(int argc, char **argv)
+{
+	 int i;
+	 uninorm_t nf = UNINORM_NFKC;
+
+	 for (i = 1; i < argc; ++i) {
+		  if (!strcmp(argv[i], "-nfkc")) {
+			   nf = UNINORM_NFKC;
+			   continue;
+		  }
+		  if (!strcmp(argv[i], "-nfkd")) {
+			   nf = UNINORM_NFKD;
+			   continue;
+		  }
+		  if (!strcmp(argv[i], "-nfc")) {
+			   nf = UNINORM_NFC;
+			   continue;
+		  }
+		  if (!strcmp(argv[i], "-nfd")) {
+			   nf = UNINORM_NFD;
+			   continue;
+		  }
+		  if (argv[i][0] == '-') {
+			   fprintf(stderr, "unrecognized option: %s\n", argv[i]);
+			   return EXIT_FAILURE;
+		  }
+
+		  size_t len;
+		  uint8_t *src = readfile(argv[i], &len);
+		  if (!src) {
+			   fprintf(stderr, "error reading %s\n", argv[i]);
+			   return EXIT_FAILURE;
+		  }
+
+		  size_t destlen;
+		  uint8_t *dest;
+		  mytime start = gettime();
+		  for (int i = 0; i < 100; ++i) {
+			   dest = u8_normalize(nf, src, len, NULL, &destlen);
+			   if (!dest) return EXIT_FAILURE;
+			   free(dest);
+		  }
+		  printf("%s: %g\n", argv[i], elapsed(gettime(), start) / 100);
+		  free(src);
+	 }
+
+	 return EXIT_SUCCESS;
+}