Add tests for valid codepoints and iterate function

.gitignore

@@ -21,3 +21,5 @@ graphemetest
 utf8proc_data.c.new
 printproperty
 charwidth
+valid
+iterate

Makefile

@@ -39,7 +39,7 @@ includedir=$(prefix)/include
 all: libutf8proc.a libutf8proc.$(SHLIB_EXT)
 
 clean:
-	rm -f utf8proc.o libutf8proc.a libutf8proc.$(SHLIB_VERS_EXT) libutf8proc.$(SHLIB_EXT) test/normtest test/graphemetest test/printproperty test/charwidth
+	rm -f utf8proc.o libutf8proc.a libutf8proc.$(SHLIB_VERS_EXT) libutf8proc.$(SHLIB_EXT) test/normtest test/graphemetest test/printproperty test/charwidth test/valid test/iterate
 	$(MAKE) -C bench clean
 	$(MAKE) -C data clean
 
@@ -105,8 +105,16 @@ test/printproperty: test/printproperty.c utf8proc.o utf8proc.h test/tests.h
 test/charwidth: test/charwidth.c utf8proc.o utf8proc.h test/tests.h
 	$(cc) test/charwidth.c utf8proc.o -o $@
 
-check: test/normtest data/NormalizationTest.txt test/graphemetest data/GraphemeBreakTest.txt test/printproperty test/charwidth bench/bench.c bench/util.c bench/util.h utf8proc.o
+test/valid: test/valid.c utf8proc.o utf8proc.h test/tests.h
+	$(cc) test/valid.c utf8proc.o -o $@
+
+test/iterate: test/iterate.c utf8proc.o utf8proc.h test/tests.h
+	$(cc) test/iterate.c utf8proc.o -o $@
+
+check: test/normtest data/NormalizationTest.txt test/graphemetest data/GraphemeBreakTest.txt test/printproperty test/charwidth test/valid test/iterate bench/bench.c bench/util.c bench/util.h utf8proc.o
 	$(MAKE) -C bench
 	test/normtest data/NormalizationTest.txt
 	test/graphemetest data/GraphemeBreakTest.txt
 	test/charwidth
+	test/valid
+	test/iterate

test/iterate.c

@@ -0,0 +1,160 @@
+#include "tests.h"
+#include <ctype.h>
+#include <wchar.h>
+
+static  int     tests;
+static  int     error;
+
+#define CHECKVALID(pos, val, len) buf[pos] = val; testbytes(buf,len,len,__LINE__)
+#define CHECKINVALID(pos, val, len) buf[pos] = val; testbytes(buf,len,UTF8PROC_ERROR_INVALIDUTF8,__LINE__)
+
+void testbytes(unsigned char *buf, int len, utf8proc_ssize_t retval, int line)
+{
+    utf8proc_int32_t out[16];
+    utf8proc_ssize_t ret;
+
+    tests++;
+    if ((ret = utf8proc_iterate(buf, len, out)) != retval) {
+        fprintf(stderr, "Failed (%d):", line);
+        for (int i = 0; i < len ; i++) {
+            fprintf(stderr, " 0x%02x", buf[i]);
+        }
+        fprintf(stderr, " -> %zd\n", ret);
+        error++;
+    }
+}
+
+int main(int argc, char **argv)
+{
+    uint32_t byt;
+    unsigned char buf[16];
+
+    tests = error = 0;
+
+    // Check valid sequences that were considered valid erroneously before
+    buf[0] = 0xef;
+    buf[1] = 0xb7;
+    for (byt = 0x90; byt < 0xa0; byt++) {
+        CHECKVALID(2, byt, 3);
+    }
+    // Check 0xfffe and 0xffff
+    buf[1] = 0xbf;
+    CHECKVALID(2, 0xbe, 3);
+    CHECKVALID(2, 0xbf, 3);
+    // Check 0x??fffe & 0x??ffff
+    for (byt = 0x1fffe; byt < 0x110000; byt += 0x10000) {
+        buf[0] = 0xf0 | (byt >> 18);
+        buf[1] = 0x80 | ((byt >> 12) & 0x3f);
+        CHECKVALID(3, 0xbe, 4);
+        CHECKVALID(3, 0xbf, 4);
+    }
+    
+    // Continuation byte not after lead
+    for (byt = 0x80; byt < 0xc0; byt++) {
+        CHECKINVALID(0, byt, 1);
+    }
+
+    // Continuation byte not after lead
+    for (byt = 0x80; byt < 0xc0; byt++) {
+        CHECKINVALID(0, byt, 1);
+    }
+
+    // Test lead bytes
+    for (byt = 0xc0; byt <= 0xff; byt++) {
+	// Single lead byte at end of string
+        CHECKINVALID(0, byt, 1);
+        // Lead followed by non-continuation character < 0x80
+        CHECKINVALID(1, 65, 2);
+	// Lead followed by non-continuation character > 0xbf
+        CHECKINVALID(1, 0xc0, 2);
+    }
+
+    // Test overlong 2-byte
+    buf[0] = 0xc0;
+    for (byt = 0x81; byt <= 0xbf; byt++) {
+        CHECKINVALID(1, byt, 2);
+    }
+    buf[0] = 0xc1;
+    for (byt = 0x80; byt <= 0xbf; byt++) {
+        CHECKINVALID(1, byt, 2);
+    }
+
+    // Test overlong 3-byte
+    buf[0] = 0xe0;
+    buf[2] = 0x80;
+    for (byt = 0x80; byt <= 0x9f; byt++) {
+        CHECKINVALID(1, byt, 3);
+    }
+
+    // Test overlong 4-byte
+    buf[0] = 0xf0;
+    buf[2] = 0x80;
+    buf[3] = 0x80;
+    for (byt = 0x80; byt <= 0x8f; byt++) {
+        CHECKINVALID(1, byt, 4);
+    }
+
+    // Test 4-byte > 0x10ffff
+    buf[0] = 0xf4;
+    buf[2] = 0x80;
+    buf[3] = 0x80;
+    for (byt = 0x90; byt <= 0xbf; byt++) {
+        CHECKINVALID(1, byt, 4);
+    }
+    buf[1] = 0x80;
+    for (byt = 0xf5; byt <= 0xf7; byt++) {
+        CHECKINVALID(0, byt, 4);
+    }
+
+    // Test 5-byte
+    buf[4] = 0x80;
+    for (byt = 0xf8; byt <= 0xfb; byt++) {
+        CHECKINVALID(0, byt, 5);
+    }
+
+    // Test 6-byte
+    buf[5] = 0x80;
+    for (byt = 0xfc; byt <= 0xfd; byt++) {
+        CHECKINVALID(0, byt, 6);
+    }
+
+    // Test 7-byte
+    buf[6] = 0x80;
+    CHECKINVALID(0, 0xfe, 7);
+
+    // Three and above byte sequences
+    for (byt = 0xe0; byt < 0xf0; byt++) {
+        // Lead followed by only 1 continuation byte
+        CHECKINVALID(0, byt, 2);
+        // Lead ended by non-continuation character < 0x80
+        CHECKINVALID(2, 65, 3);
+        // Lead ended by non-continuation character > 0xbf
+        CHECKINVALID(2, 0xc0, 3);
+    }
+
+    // 3-byte encoded surrogate character(s)
+    buf[0] = 0xed; buf[2] = 0x80;
+    // Single surrogate
+    CHECKINVALID(1, 0xa0, 3);
+    // Trailing surrogate first
+    CHECKINVALID(1, 0xb0, 3);
+
+    // Four byte sequences
+    buf[1] = 0x80;
+    for (byt = 0xf0; byt < 0xf5; byt++) {
+        // Lead followed by only 1 continuation bytes
+        CHECKINVALID(0, byt, 2);
+        // Lead followed by only 2 continuation bytes
+        CHECKINVALID(0, byt, 3);
+        // Lead followed by non-continuation character < 0x80
+        CHECKINVALID(3, 65, 4);
+        // Lead followed by non-continuation character > 0xbf
+        CHECKINVALID(3, 0xc0, 4);
+
+    }
+
+     check(!error, "utf8proc_iterate FAILED %d tests out of %d", error, tests);
+     printf("utf8proc_iterate tests SUCCEEDED, (%d) tests passed.\n", tests);
+
+     return 0;
+}

test/valid.c

@@ -0,0 +1,41 @@
+#include "tests.h"
+#include <ctype.h>
+#include <wchar.h>
+
+int main(int argc, char **argv)
+{
+     int c, error = 0;
+
+     (void) argc; /* unused */
+     (void) argv; /* unused */
+
+     /* some simple sanity tests of  */
+     for (c = 0; c < 0xd800; c++) {
+        if (!utf8proc_codepoint_valid(c)) {
+           fprintf(stderr, "Failed: codepoint_valid(%04x) -> false\n", c);
+           error++;
+        }
+     }
+     for (;c < 0xe000; c++) {
+        if (utf8proc_codepoint_valid(c)) {
+           fprintf(stderr, "Failed: codepoint_valid(%04x) -> true\n", c);
+           error++;
+        }
+     }
+     for (;c < 0x110000; c++) {
+        if (!utf8proc_codepoint_valid(c)) {
+           fprintf(stderr, "Failed: codepoint_valid(%06x) -> false\n", c);
+           error++;
+        }
+     }
+     for (;c < 0x110010; c++) {
+        if (utf8proc_codepoint_valid(c)) {
+           fprintf(stderr, "Failed: codepoint_valid(%06x) -> true\n", c);
+           error++;
+        }
+     }
+     check(!error, "utf8proc_codepoint_valid FAILED %d tests.", error);
+     printf("Validity tests SUCCEEDED.\n");
+
+     return 0;
+}