From 6249e6b8b1d5e360b5b5adad651911f3b4359604 Mon Sep 17 00:00:00 2001 From: Scott Paul Jones Date: Sat, 9 May 2015 11:45:03 -0400 Subject: Fix #34 handle 66 Unicode non-characters, also improve performance and surrogate handling --- test/graphemetest.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'test/graphemetest.c') diff --git a/test/graphemetest.c b/test/graphemetest.c index a05b6e2..eb3645b 100644 --- a/test/graphemetest.c +++ b/test/graphemetest.c @@ -6,6 +6,7 @@ int main(int argc, char **argv) size_t bufsize = 0; FILE *f = argc > 1 ? fopen(argv[1], "r") : NULL; utf8proc_uint8_t src[1024]; + int len; check(f != NULL, "error opening GraphemeBreakTest.txt"); while (getline(&buf, &bufsize, f) > 0) { @@ -29,9 +30,10 @@ int main(int argc, char **argv) else if (buf[bi] == '#') { /* start of comments */ break; } - else { /* hex-encoded codepoint */ - bi += encode((char*) (src + si), buf + bi) - 1; + else { /* hex-encoded codepoint */ + len = encode((char*) (src + si), buf + bi) - 1; while (src[si]) ++si; /* advance to NUL termination */ + bi += len; } } if (si && src[si-1] == '/') @@ -59,7 +61,7 @@ int main(int argc, char **argv) utf8proc_errmsg(glen)); for (i = 0; i <= glen; ++i) if (g[i] == 0xff) - g[i] = '/'; /* easier-to-read output (/ is not in test strings) */ + g[i] = '/'; /* easier-to-read output (/ is not in test strings) */ check(!strcmp((char*)g, (char*)src), "grapheme mismatch: \"%s\" instead of \"%s\"", (char*)g, (char*)src); } -- cgit v1.2.3