From b3e70932b6534e603b03d28c45133b8f991f48fe Mon Sep 17 00:00:00 2001 From: Rob Landley Date: Tue, 5 Sep 2017 02:36:24 -0500 Subject: Tweak utf8towc() to return -1 earlier sometimes (instead of -2), and add test program to compare against libc output. --- lib/lib.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) (limited to 'lib/lib.c') diff --git a/lib/lib.c b/lib/lib.c index c482dcab..a4b7229b 100644 --- a/lib/lib.c +++ b/lib/lib.c @@ -345,20 +345,17 @@ int utf8towc(wchar_t *wc, char *str, unsigned len) if (len && *str<128) return !!(*wc = *str); result = first = *(s = str++); + if (result<0xc2 || result>0xf4) return -1; for (mask = 6; (first&0xc0)==0xc0; mask += 5, first <<= 1) { - if (mask>21) return -1; if (!--len) return -2; - c = *(str++); - if ((c&0xc0) != 0x80) return -1; + if (((c = *(str++))&0xc0) != 0x80) return -1; result = (result<<6)|(c&0x3f); } result &= (1<21 || result<(unsigned []){0x80,0x800,0x10000}[c-2]) - return -1; + if (result<(unsigned []){0x80,0x800,0x10000}[c-2]) return -1; // Limit unicode so it can't encode anything UTF-16 can't. if (result>0x10ffff || (result>=0xd800 && result<=0xdfff)) return -1; -- cgit v1.2.3