Merge branch 'encode-speedup' of https://github.com/bpowers/yajl into 2.1.0

1 files changed, 35 insertions, 13 deletions

diff --git a/src/yajl_encode.c b/src/yajl_encode.c
index 9dc9a3e..9770df0 100644
--- a/src/yajl_encode.c
+++ b/src/yajl_encode.c
@@ -23,7 +23,7 @@
 
 static void CharToHex(unsigned char c, char * hexBuf)
 {
-    const char * hexchar = "0123456789ABCDEF";
+    static const char * hexchar = "0123456789ABCDEF";
     hexBuf[0] = hexchar[c >> 4];
     hexBuf[1] = hexchar[c & 0x0F];
 }
@@ -37,36 +37,58 @@ yajl_string_encode(const yajl_print_t print,
 {
     size_t beg = 0;
     size_t end = 0;
+    size_t escaped_len;
+    char escBuf[3];
     char hexBuf[7];
+    escBuf[0] = '\\';
+    escBuf[2] = 0;
     hexBuf[0] = '\\'; hexBuf[1] = 'u'; hexBuf[2] = '0'; hexBuf[3] = '0';
     hexBuf[6] = 0;
 
     while (end < len) {
+        const unsigned char chr = str[end];
         const char * escaped = NULL;
-        switch (str[end]) {
-            case '\r': escaped = "\\r"; break;
-            case '\n': escaped = "\\n"; break;
-            case '\\': escaped = "\\\\"; break;
+        escBuf[1] = 0;
+        /* we're not going to be escaping most characters, so first
+         * check for this common case (of doing nothing).  Doing so
+         * decreases the runtime of this function by around 30%.
+         */
+        if (chr > '/') {
             /* it is not required to escape a solidus in JSON:
              * read sec. 2.5: http://www.ietf.org/rfc/rfc4627.txt
              * specifically, this production from the grammar:
              *   unescaped = %x20-21 / %x23-5B / %x5D-10FFFF
              */
-            case '/': if (escape_solidus) escaped = "\\/"; break;
-            case '"': escaped = "\\\""; break;
-            case '\f': escaped = "\\f"; break;
-            case '\b': escaped = "\\b"; break;
-            case '\t': escaped = "\\t"; break;
+            if (chr == '\\') {
+                escBuf[1] = '\\';
+            }
+        } else if (chr == ' ' || (chr <  '/' && chr > ')')) {
+            /* skip these fairly common cases */
+        } else {
+            switch (chr) {
+            case '\r': escBuf[1] = 'r'; break;
+            case '\n': escBuf[1] = 'n'; break;
+            case '"':  escBuf[1] = '"'; break;
+            case '\t': escBuf[1] = 't'; break;
+            case '\f': escBuf[1] = 'f'; break;
+            case '\b': escBuf[1] = 'b'; break;
+            case '/': if (escape_solidus) escBuf[1] = '/'; break;
             default:
-                if ((unsigned char) str[end] < 32) {
-                    CharToHex(str[end], hexBuf + 4);
+                if (chr < 32) {
+                    CharToHex(chr, hexBuf + 4);
                     escaped = hexBuf;
+                    escaped_len = 6;
                 }
                 break;
+            }
+        }
+        if (escBuf[1] != 0) {
+                escaped = escBuf;
+                escaped_len = 2;
         }
         if (escaped != NULL) {
             print(ctx, (const char *) (str + beg), end - beg);
-            print(ctx, escaped, (unsigned int)strlen(escaped));
+            print(ctx, escaped, escaped_len);
             beg = ++end;
         } else {
             ++end;