From 9b3d1f893ac0481b7024bd04c11d5f2c82a2d90c Mon Sep 17 00:00:00 2001
From: Jean Boussier <jean.boussier@gmail.com>
Date: Mon, 20 Jun 2022 20:35:50 +0200
Subject: [PATCH] Implement `:escape_mode => :slash`

Similar to https://github.com/flori/json/pull/405
It's a cheap way to make JSON safe to interpolate in a `<script>`
tag.
---
 ext/oj/dump.c        | 21 +++++++++++++++++++++
 ext/oj/oj.c          |  6 ++++++
 ext/oj/oj.h          |  1 +
 test/test_various.rb |  6 ++++++
 4 files changed, 34 insertions(+)

diff --git a/ext/oj/dump.c b/ext/oj/dump.c
index 83f4add5..ed7340cd 100644
--- a/ext/oj/dump.c
+++ b/ext/oj/dump.c
@@ -32,6 +32,7 @@ static const char nan_val[]  = NAN_VAL;
 typedef unsigned long ulong;
 
 static size_t hibit_friendly_size(const uint8_t *str, size_t len);
+static size_t slash_friendly_size(const uint8_t *str, size_t len);
 static size_t xss_friendly_size(const uint8_t *str, size_t len);
 static size_t ascii_friendly_size(const uint8_t *str, size_t len);
 
@@ -59,6 +60,17 @@ static char hibit_friendly_chars[256] = "\
 11111111111111111111111111111111\
 11111111111111111111111111111111";
 
+// JSON standard but escape forward slashes `/`
+static char slash_friendly_chars[256] = "\
+66666666222622666666666666666666\
+11211111111111121111111111111111\
+11111111111111111111111111112111\
+11111111111111111111111111111111\
+11111111111111111111111111111111\
+11111111111111111111111111111111\
+11111111111111111111111111111111\
+11111111111111111111111111111111";
+
 // High bit set characters are always encoded as unicode. Worse case is 3
 // bytes per character in the output. That makes this conservative.
 static char ascii_friendly_chars[256] = "\
@@ -143,6 +155,10 @@ inline static size_t hibit_friendly_size(const uint8_t *str, size_t len) {
     return calculate_string_size(str, len, hibit_friendly_chars);
 }
 
+inline static size_t slash_friendly_size(const uint8_t *str, size_t len) {
+    return calculate_string_size(str, len, slash_friendly_chars);
+}
+
 inline static size_t ascii_friendly_size(const uint8_t *str, size_t len) {
     return calculate_string_size(str, len, ascii_friendly_chars);
 }
@@ -756,6 +772,11 @@ void oj_dump_cstr(const char *str, size_t cnt, bool is_sym, bool escape1, Out ou
         cmap = ascii_friendly_chars;
         size = ascii_friendly_size((uint8_t *)str, cnt);
         break;
+    case SlashEsc:
+        has_hi = true;
+        cmap = slash_friendly_chars;
+        size = slash_friendly_size((uint8_t *)str, cnt);
+        break;
     case XSSEsc:
         cmap = xss_friendly_chars;
         size = xss_friendly_size((uint8_t *)str, cnt);
diff --git a/ext/oj/oj.c b/ext/oj/oj.c
index 0df10104..b9634d27 100644
--- a/ext/oj/oj.c
+++ b/ext/oj/oj.c
@@ -137,6 +137,7 @@ static VALUE rails_sym;
 static VALUE raise_sym;
 static VALUE ruby_sym;
 static VALUE sec_prec_sym;
+static VALUE slash_sym;
 static VALUE strict_sym;
 static VALUE symbol_keys_sym;
 static VALUE time_format_sym;
@@ -405,6 +406,7 @@ static VALUE get_def_opts(VALUE self) {
     switch (oj_default_options.escape_mode) {
     case NLEsc: rb_hash_aset(opts, escape_mode_sym, newline_sym); break;
     case JSONEsc: rb_hash_aset(opts, escape_mode_sym, json_sym); break;
+    case SlashEsc: rb_hash_aset(opts, escape_mode_sym, slash_sym); break;
     case XSSEsc: rb_hash_aset(opts, escape_mode_sym, xss_safe_sym); break;
     case ASCIIEsc: rb_hash_aset(opts, escape_mode_sym, ascii_sym); break;
     case JXEsc: rb_hash_aset(opts, escape_mode_sym, unicode_xss_sym); break;
@@ -734,6 +736,8 @@ static int parse_options_cb(VALUE k, VALUE v, VALUE opts) {
             copts->escape_mode = NLEsc;
         } else if (json_sym == v) {
             copts->escape_mode = JSONEsc;
+        } else if (slash_sym == v) {
+            copts->escape_mode = SlashEsc;
         } else if (xss_safe_sym == v) {
             copts->escape_mode = XSSEsc;
         } else if (ascii_sym == v) {
@@ -1987,6 +1991,8 @@ void Init_oj(void) {
     rb_gc_register_address(&ruby_sym);
     sec_prec_sym = ID2SYM(rb_intern("second_precision"));
     rb_gc_register_address(&sec_prec_sym);
+    slash_sym = ID2SYM(rb_intern("slash"));
+    rb_gc_register_address(&slash_sym);
     strict_sym = ID2SYM(rb_intern("strict"));
     rb_gc_register_address(&strict_sym);
     symbol_keys_sym = ID2SYM(rb_intern("symbol_keys"));
diff --git a/ext/oj/oj.h b/ext/oj/oj.h
index 3895e828..9eb8b97f 100644
--- a/ext/oj/oj.h
+++ b/ext/oj/oj.h
@@ -66,6 +66,7 @@ typedef enum { UnixTime = 'u', UnixZTime = 'z', XmlTime = 'x', RubyTime = 'r' }
 typedef enum {
     NLEsc     = 'n',
     JSONEsc   = 'j',
+    SlashEsc  = 's',
     XSSEsc    = 'x',
     ASCIIEsc  = 'a',
     JXEsc     = 'g',  // json gem
diff --git a/test/test_various.rb b/test/test_various.rb
index d1a92e75..d6bea7b6 100755
--- a/test/test_various.rb
+++ b/test/test_various.rb
@@ -345,6 +345,12 @@ def test_escapes_entities_by_default_when_configured_to_do_so
     out = Oj.dump hash
     assert_equal(%{{"key":"I \\u003c3 this"}}, out)
   end
+  def test_escapes_slashes_by_default_when_configured_to_do_so
+    hash = {'key' => "I <3 this </script>"}
+    Oj.default_options = {:escape_mode => :slash}
+    out = Oj.dump hash
+    assert_equal(%{{"key":"I <3 this <\\/script>"}}, out)
+  end
   def test_escapes_entities_when_asked_to
     hash = {'key' => "I <3 this"}
     out = Oj.dump(hash, :escape_mode => :xss_safe)