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.

1 files changed, 70 insertions, 0 deletions

diff --git a/sys/base/arg.c b/sys/base/arg.c
index 64e4dd6..269043e 100644
--- a/sys/base/arg.c
+++ b/sys/base/arg.c
@@ -1 +1,71 @@
+#include <u.h>
+#include <base.h>
+
+// NOTE: this utf8 bit is copied from libunicode to remove the hard dependency just for ARG_BEGIN.
+
+#define UTFmax   4
+#define RuneSync 0x80u
+#define RuneSelf 0x80u
+#define RuneErr  0xFFFDu
+#define RuneMax  0x10FFFFu
+#define RuneMask 0x1FFFFFu
+
+#define Bit(i) (7-(i))
+/* N 0's preceded by i 1's e.g. T(Bit(2)) is 1100 0000 */
+#define Tbyte(i) (((1 << (Bit(i)+1))-1) ^ 0xFF)
+/* 0000 0000 0000 0111 1111 1111 */
+#define	RuneX(i) ((1 << (Bit(i) + ((i)-1)*Bitx))-1)
+enum
+{
+    Bitx  = Bit(1),
+    Tx    = Tbyte(1),
+    Rune1 = (1 << (Bit(0)+0*Bitx)) - 1,
+
+    Maskx = (1 << Bitx) - 1, /* 0011 1111 */
+    Testx = Maskx ^ 0xff,    /* 1100 0000 */
+
+    SurrogateMin = 0xD800,
+    SurrogateMax = 0xDFFF,
+    Bad = RuneErr,
+};
+
+
+int
+arg·bytetorune(uint32* r, byte* s)
+{
+    int c[4], i;
+    uint32 l;
+
+    c[0] = *(ubyte*)(s);
+    if(c[0] < Tx) {
+        *r = c[0];
+        return 1;
+    }
+
+    l = c[0];
+    for(i = 1; i < UTFmax; i++) {
+        c[i]  = *(ubyte*)(s+i);
+        c[i] ^= Tx;
+        if (c[i] & Testx) goto bad;
+
+        l = (l << Bitx) | c[i];
+        if(c[0] < Tbyte(i + 2)) {
+            l &= RuneX(i + 1);
+            if (i == 1) {
+                if (c[0] < Tbyte(2) || l <= Rune1)
+                    goto bad;
+            } else if (l <= RuneX(i) || l > RuneMax)
+                goto bad;
+            if (i == 2 && SurrogateMin <= l && l <= SurrogateMax)
+                goto bad;
+
+            *r = l;
+            return i + 1;
+        }
+    }
+bad:
+    *r = RuneErr;
+    return 1;
+}
+
 char *argv0;