Add unescape2(), migrate some unescape() users over.

4 files changed, 38 insertions, 48 deletions

diff --git a/lib/lib.c b/lib/lib.c
index 364ce07e..4fa358c0 100644
--- a/lib/lib.c
+++ b/lib/lib.c
@@ -436,6 +436,29 @@ int unescape(char c)
   return (idx == -1) ? 0 : to[idx];
 }
 
+// parse next character advancing pointer. echo requires leading 0 in octal esc
+int unescape2(char **c, int echo)
+{
+  int idx = *((*c)++), i, off;
+
+  if (idx != '\\' || !**c) return idx;
+  if (**c == 'c') return 31&*(++*c);
+  for (i = 0; i<4; i++) {
+    if (sscanf(*c, (char *[]){"0%3o%n"+!echo, "x%2x%n", "u%4x%n", "U%6x%n"}[i],
+        &idx, &off))
+    {
+      *c += off;
+
+      return idx;
+    }
+  }
+
+  if (-1 == (idx = stridx("\\abeEfnrtv'\"?", **c))) return '\\';
+  ++*c;
+
+  return "\\\a\b\033\033\f\n\r\t\v'\"?"[idx];
+}
+
 // If string ends with suffix return pointer to start of suffix in string,
 // else NULL
 char *strend(char *str, char *suffix)
diff --git a/lib/lib.h b/lib/lib.h
index 3200dc32..60c800ca 100644
--- a/lib/lib.h
+++ b/lib/lib.h
@@ -234,6 +234,7 @@ char *strlower(char *s);
 char *strafter(char *haystack, char *needle);
 char *chomp(char *s);
 int unescape(char c);
+int unescape2(char **c, int echo);
 char *strend(char *str, char *suffix);
 int strstart(char **a, char *b);
 int strcasestart(char **a, char *b);
diff --git a/toys/posix/echo.c b/toys/posix/echo.c
index c78c2c01..b546e94b 100644
--- a/toys/posix/echo.c
+++ b/toys/posix/echo.c
@@ -41,12 +41,10 @@ config ECHO
 
 void echo_main(void)
 {
-  int i = 0, out;
-  char *arg, *c;
+  int i = 0;
+  char *arg, *c, out[8];
 
-  for (;;) {
-    arg = toys.optargs[i];
-    if (!arg) break;
+  while ((arg = toys.optargs[i])) {
     if (i++) putchar(' ');
 
     // Should we output arg verbatim?
@@ -58,40 +56,12 @@ void echo_main(void)
 
     // Handle -e
 
-    for (c = arg;;) {
-      if (!(out = *(c++))) break;
+    for (c = arg; *c; ) {
+      unsigned u;
 
-      // handle \escapes
-      if (out == '\\' && *c) {
-        int slash = *(c++), n = unescape(slash);
-
-        if (n) out = n;
-        else if (slash=='c') return;
-        else if (slash=='0') {
-          out = 0;
-          while (*c>='0' && *c<='7' && n++<3) out = (out*8)+*(c++)-'0';
-        } else if (slash=='x') {
-          out = 0;
-          while (n++<2) {
-            if (*c>='0' && *c<='9') out = (out*16)+*(c++)-'0';
-            else {
-              int temp = tolower(*c);
-              if (temp>='a' && temp<='f') {
-                out = (out*16)+temp-'a'+10;
-                c++;
-              } else {
-                if (n==1) {
-                  --c;
-                  out = '\\';
-                }
-                break;
-              }
-            }
-          }
-        // Slash in front of unknown character, print literal.
-        } else c--;
-      }
-      putchar(out);
+      if (*c == '\\' && c[1] == 'c') return;
+      if ((u = unescape2(&c, 1))<128) putchar(u);
+      else printf("%.*s", (int)wcrtomb(out, u, 0), out);
     }
   }
 
diff --git a/toys/posix/find.c b/toys/posix/find.c
index 3fb97b1c..9b688d46 100644
--- a/toys/posix/find.c
+++ b/toys/posix/find.c
@@ -585,16 +585,12 @@ static int do_find(struct dirtree *new)
         if (check) for (fmt = ss[1]; *fmt; fmt++) {
           // Print the parts that aren't escapes
           if (*fmt == '\\') {
-            int slash = *++fmt, n = unescape(slash);
-
-            if (n) ch = n;
-            else if (slash=='c') break;
-            else if (slash=='0') {
-              ch = 0;
-              while (*fmt>='0' && *fmt<='7' && n++<3) ch=(ch*8)+*(fmt++)-'0';
-              --fmt;
-            } else error_exit("bad \\%c", *fmt);
-            putchar(ch);
+            unsigned u;
+
+            if (fmt[1] == 'c') break;
+            if ((u = unescape2(&fmt, 0))<128) putchar(u);
+            else printf("%.*s", (int)wcrtomb(buf, u, 0), buf);
+            fmt--;
           } else if (*fmt != '%') putchar(*fmt);
           else if (*++fmt == '%') putchar('%');
           else {