aboutsummaryrefslogtreecommitdiff
path: root/sys/libunicode/runetobyte.c
diff options
context:
space:
mode:
authorNicholas Noll <nbnoll@eml.cc>2021-10-26 21:01:41 -0700
committerNicholas Noll <nbnoll@eml.cc>2021-10-26 21:01:41 -0700
commit29b56ef4e4113bcd091b19d6926f18814162ca53 (patch)
treea4888a16927576592af13928bb805f1f2b1159d6 /sys/libunicode/runetobyte.c
parente34a4791b72e426b02f33496fe03be1ad81819a6 (diff)
Feat(libunicode): Added an explicit unicode library
Refactored code to pull out utf8 functions from base into a standalone library. Also left the required function inside arg.c so that code that calls ARG_BEGIN doesn't have to link to libunicode.
Diffstat (limited to 'sys/libunicode/runetobyte.c')
-rw-r--r--sys/libunicode/runetobyte.c31
1 files changed, 31 insertions, 0 deletions
diff --git a/sys/libunicode/runetobyte.c b/sys/libunicode/runetobyte.c
new file mode 100644
index 0000000..27f252b
--- /dev/null
+++ b/sys/libunicode/runetobyte.c
@@ -0,0 +1,31 @@
+#include "internal.h"
+
+int
+utf8·runetobyte(byte* s, rune* r)
+{
+ int i, j;
+ rune c;
+
+ c = *r;
+ if(c <= Rune1) {
+ s[0] = c;
+ return 1;
+ }
+
+ for(i = 2; i < UTFmax + 1; i++){
+ if(i == 3){
+ if(c > RuneMax)
+ c = RuneErr;
+ if(SurrogateMin <= c && c <= SurrogateMax)
+ c = RuneErr;
+ }
+ if(c <= RuneX(i) || i == UTFmax) {
+ s[0] = Tbyte(i) | (c >> (i - 1)*Bitx);
+ for(j = 1; j < i; j++)
+ s[j] = Tx | ((c >> (i - j - 1)*Bitx) & Maskx);
+ return i;
+ }
+ }
+
+ return UTFmax;
+}