cgi/escape: Optimize CGI.escapeHTML

* cgi/escape/escape.c: Optimize CGI.escapeHTML for
  ASCII-compatible encodings.  [Fix GH-1164]

git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@53220 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

2 files changed, 100 insertions, 0 deletions

diff --git a/ext/cgi/escape/escape.c b/ext/cgi/escape/escape.c
new file mode 100644
index 0000000000..6fec95af04
--- /dev/null
+++ b/ext/cgi/escape/escape.c
@@ -0,0 +1,97 @@
+#include "ruby.h"
+#include "ruby/encoding.h"
+
+static VALUE rb_cCGI, rb_mUtil, rb_mEscape;
+
+static void
+html_escaped_cat(VALUE str, char c)
+{
+    switch (c) {
+      case '\'':
+	rb_str_cat_cstr(str, "'");
+	break;
+      case '&':
+	rb_str_cat_cstr(str, "&");
+	break;
+      case '"':
+	rb_str_cat_cstr(str, """);
+	break;
+      case '<':
+	rb_str_cat_cstr(str, "&lt;");
+	break;
+      case '>':
+	rb_str_cat_cstr(str, "&gt;");
+	break;
+    }
+}
+
+static VALUE
+optimized_escape_html(VALUE str)
+{
+    long i, len, modified = 0, beg = 0;
+    VALUE dest;
+    const char *cstr;
+
+    len  = RSTRING_LEN(str);
+    cstr = RSTRING_PTR(str);
+
+    for (i = 0; i < len; i++) {
+	switch (cstr[i]) {
+	  case '\'':
+	  case '&':
+	  case '"':
+	  case '<':
+	  case '>':
+	    if (!modified) {
+		modified = 1;
+		dest = rb_str_buf_new(len);
+	    }
+
+	    rb_str_cat(dest, cstr + beg, i - beg);
+	    beg = i + 1;
+
+	    html_escaped_cat(dest, cstr[i]);
+	    break;
+	}
+    }
+
+    if (modified) {
+	rb_str_cat(dest, cstr + beg, len - beg);
+	rb_enc_associate(dest, rb_enc_get(str));
+	return dest;
+    }
+    else {
+	return str;
+    }
+}
+
+/*
+ *  call-seq:
+ *     CGI.escapeHTML(string) -> string
+ *
+ *  Returns HTML-escaped string.
+ *
+ */
+static VALUE
+cgiesc_escape_html(VALUE self, VALUE str)
+{
+    StringValue(str);
+
+    if (rb_enc_str_asciicompat_p(str)) {
+	return optimized_escape_html(str);
+    }
+    else {
+	return rb_call_super(1, &str);
+    }
+}
+
+void
+Init_escape(void)
+{
+    rb_cCGI    = rb_define_class("CGI", rb_cObject);
+    rb_mEscape = rb_define_module_under(rb_cCGI, "Escape");
+    rb_mUtil   = rb_define_module_under(rb_cCGI, "Util");
+    rb_define_method(rb_mEscape, "escapeHTML", cgiesc_escape_html, 1);
+    rb_prepend_module(rb_mUtil, rb_mEscape);
+    rb_extend_object(rb_cCGI, rb_mEscape);
+}
diff --git a/ext/cgi/escape/extconf.rb b/ext/cgi/escape/extconf.rb
new file mode 100644
index 0000000000..16e8ff224d
--- /dev/null
+++ b/ext/cgi/escape/extconf.rb
@@ -0,0 +1,3 @@
+require 'mkmf'
+
+create_makefile 'cgi/escape'