bson 4.8.1 → 4.9.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: ef5a31f0820c55727982451d554e05971480a6df19a3a959eea4731326193614
-  data.tar.gz: 8928dec44431fa67fba6f5b95a469ee5e8c8d17d12880ac5c35aa556acff697a
+  metadata.gz: 1d24a5aa6f5678defced03d887f8d946a7c03992cff980281138bb5836b83366
+  data.tar.gz: b22e4544322858b2473b83f4842ba7074efd9a2183c181818ffe7120ae85bf1d
 SHA512:
-  metadata.gz: 48e4afe1f4da2dd9161e326f05c9c3dcf0f287845eb46fb3eebfe08e761df0e73d36bf0745867fca70ace7345d8a243826da9a0f89f9b3c4a28e903b74817949
-  data.tar.gz: b40043516efec6952d2f83ffa84183ed4e7e23a545ed9b51e12729d65f6a944131468ea4278dccf4cdde5e6ec9f859bc9cad04aeae167f9a143ebe62731df9e3
+  metadata.gz: f54379785c0d7c5dbd9e75919aa8ab5f0b6fa04158a2f4789c47bf763b763901c0a64255f4aadffc62a76cc3068e57d31352a49e5acc7fd5e6766fe7d7a32b3b
+  data.tar.gz: 667307d98f30bfc0ea6a3a15bdfed67fd9a935c4512504bae9a2f545d2e1730f1105facf4acd11d03a315efacb877d531029feec01b77d7a54d1182df297b8ae

data/README.md

@@ -1,8 +1,9 @@
-BSON [![Build Status](https://secure.travis-ci.org/mongodb/bson-ruby.svg?branch=master)](http://travis-ci.org/mongodb/bson-ruby)
-[![Build status Windows](https://ci.appveyor.com/api/projects/status/p5aqko7umsx351nm?svg=true)](https://ci.appveyor.com/project/p-mongo/bson-ruby/branch/master)
-[![Code Climate](https://codeclimate.com/github/mongodb/bson-ruby.svg)](https://codeclimate.com/github/mongodb/bson-ruby)
-[![Coverage Status](https://coveralls.io/repos/mongodb/bson-ruby/badge.svg?branch=master)](https://coveralls.io/r/mongodb/bson-ruby?branch=master)
-[![Inline docs](http://inch-ci.org/github/mongodb/bson-ruby.svg?branch=master)](http://inch-ci.org/github/mongodb/bson-ruby)
+BSON
+[![Gem Version][rubygems-img]][rubygems-url]
+[![Build Status][travis-img]][travis-url]
+[![Build status Windows][appveyor-img]][appveyor-url]
+[![Coverage Status][coveralls-img]][coveralls-url]
+[![Inline docs][inch-img]][inch-url]
 ====
 
 An implementation of the BSON specification in Ruby.
@@ -38,7 +39,7 @@ As of 2.0.0, this project adheres to the
 License
 -------
 
-Copyright (C) 2009-2019 MongoDB Inc.
+Copyright (C) 2009-2020 MongoDB Inc.
 
 Licensed under the Apache License, Version 2.0 (the "License");
 you may not use this file except in compliance with the License.
@@ -51,3 +52,14 @@ distributed under the License is distributed on an "AS IS" BASIS,
 WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License.
+
+[rubygems-img]: https://badge.fury.io/rb/bson.svg
+[rubygems-url]: http://badge.fury.io/rb/bson
+[travis-img]: https://secure.travis-ci.org/mongodb/bson-ruby.svg?branch=master
+[travis-url]: http://travis-ci.org/mongodb/bson-ruby
+[appveyor-img]: https://ci.appveyor.com/api/projects/status/p5aqko7umsx351nm?svg=true
+[appveyor-url]: https://ci.appveyor.com/project/p-mongo/bson-ruby/branch/master
+[coveralls-img]: https://coveralls.io/repos/mongodb/bson-ruby/badge.svg?branch=master
+[coveralls-url]: https://coveralls.io/r/mongodb/bson-ruby?branch=master
+[inch-img]: http://inch-ci.org/github/mongodb/bson-ruby.svg?branch=master
+[inch-url]: http://inch-ci.org/github/mongodb/bson-ruby

data/ext/bson/bson-endian.h

@@ -1,5 +1,5 @@
 /*
- * Copyright (C) 2015-2019 MongoDB, Inc.
+ * Copyright (C) 2015-2020 MongoDB Inc.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.

data/ext/bson/bson-native.h

@@ -1,5 +1,5 @@
 /*
- * Copyright (C) 2009-2019 MongoDB, Inc.
+ * Copyright (C) 2009-2020 MongoDB Inc.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.

data/ext/bson/bytebuf.c

@@ -1,5 +1,5 @@
 /*
- * Copyright (C) 2009-2019 MongoDB, Inc.
+ * Copyright (C) 2009-2020 MongoDB Inc.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.

data/ext/bson/endian.c

@@ -1,5 +1,5 @@
 /*
- * Copyright (C) 2015-2019 MongoDB, Inc.
+ * Copyright (C) 2015-2020 MongoDB Inc.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.

data/ext/bson/init.c

@@ -1,5 +1,5 @@
 /*
- * Copyright (C) 2009-2019 MongoDB Inc.
+ * Copyright (C) 2009-2020 MongoDB Inc.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -38,25 +38,26 @@ void Init_bson_native()
   char rb_bson_machine_id[256];
 
   VALUE rb_bson_module = rb_define_module("BSON");
-  
+
   /* Document-class: BSON::ByteBuffer
    *
    * Stores BSON-serialized data and provides efficient serialization and
    * deserialization of common Ruby classes using native code.
    */
   VALUE rb_byte_buffer_class = rb_define_class_under(rb_bson_module, "ByteBuffer", rb_cObject);
-  
+
   VALUE rb_bson_object_id_class = rb_const_get(rb_bson_module, rb_intern("ObjectId"));
   VALUE rb_bson_object_id_generator_class = rb_const_get(rb_bson_object_id_class, rb_intern("Generator"));
   VALUE rb_digest_class = rb_const_get(rb_cObject, rb_intern("Digest"));
   VALUE rb_md5_class = rb_const_get(rb_digest_class, rb_intern("MD5"));
 
   rb_bson_illegal_key = rb_const_get(rb_const_get(rb_bson_module, rb_intern("String")),rb_intern("IllegalKey"));
+  rb_gc_register_mark_object(rb_bson_illegal_key);
 
   rb_define_alloc_func(rb_byte_buffer_class, rb_bson_byte_buffer_allocate);
   rb_define_method(rb_byte_buffer_class, "initialize", rb_bson_byte_buffer_initialize, -1);
   rb_define_method(rb_byte_buffer_class, "length", rb_bson_byte_buffer_length, 0);
-  
+
   /*
    * call-seq:
    *   buffer.read_position -> Fixnum
@@ -64,13 +65,13 @@ void Init_bson_native()
    * Returns the read position in the buffer.
    */
   rb_define_method(rb_byte_buffer_class, "read_position", rb_bson_byte_buffer_read_position, 0);
-  
+
   rb_define_method(rb_byte_buffer_class, "get_byte", rb_bson_byte_buffer_get_byte, 0);
   rb_define_method(rb_byte_buffer_class, "get_bytes", rb_bson_byte_buffer_get_bytes, 1);
   rb_define_method(rb_byte_buffer_class, "get_cstring", rb_bson_byte_buffer_get_cstring, 0);
   rb_define_method(rb_byte_buffer_class, "get_decimal128_bytes", rb_bson_byte_buffer_get_decimal128_bytes, 0);
   rb_define_method(rb_byte_buffer_class, "get_double", rb_bson_byte_buffer_get_double, 0);
-  
+
   /*
    * call-seq:
    *   buffer.get_hash(**options) -> Hash
@@ -82,7 +83,7 @@ void Init_bson_native()
    * @return [ BSON::Document ] The decoded document.
    */
   rb_define_method(rb_byte_buffer_class, "get_hash", rb_bson_byte_buffer_get_hash, -1);
-  
+
   /*
    * call-seq:
    *   buffer.get_array(**options) -> Array
@@ -94,11 +95,11 @@ void Init_bson_native()
    * @return [ Array ] The decoded array.
    */
   rb_define_method(rb_byte_buffer_class, "get_array", rb_bson_byte_buffer_get_array, -1);
-  
+
   rb_define_method(rb_byte_buffer_class, "get_int32", rb_bson_byte_buffer_get_int32, 0);
   rb_define_method(rb_byte_buffer_class, "get_int64", rb_bson_byte_buffer_get_int64, 0);
   rb_define_method(rb_byte_buffer_class, "get_string", rb_bson_byte_buffer_get_string, 0);
-  
+
   /*
    * call-seq:
    *   buffer.write_position -> Fixnum
@@ -106,7 +107,7 @@ void Init_bson_native()
    * Returns the write position in the buffer.
    */
   rb_define_method(rb_byte_buffer_class, "write_position", rb_bson_byte_buffer_write_position, 0);
-  
+
   /*
    * call-seq:
    *   buffer.put_byte(binary_str) -> ByteBuffer
@@ -117,7 +118,7 @@ void Init_bson_native()
    * Returns the modified +self+.
    */
   rb_define_method(rb_byte_buffer_class, "put_byte", rb_bson_byte_buffer_put_byte, 1);
-  
+
   /*
    * call-seq:
    *   buffer.put_bytes(binary_str) -> ByteBuffer
@@ -170,7 +171,7 @@ void Init_bson_native()
    * instead on the null terminator), unlike the BSON string serialization.
    */
   rb_define_method(rb_byte_buffer_class, "put_cstring", rb_bson_byte_buffer_put_cstring, 1);
-  
+
   /**
    * call-seq:
    *   buffer.put_symbol(sym) -> ByteBuffer
@@ -188,7 +189,7 @@ void Init_bson_native()
    * indistinguishable from a string with the same value written to the buffer.
    */
   rb_define_method(rb_byte_buffer_class, "put_symbol", rb_bson_byte_buffer_put_symbol, 1);
-  
+
   /*
    * call-seq:
    *   buffer.put_int32(fixnum) -> ByteBuffer
@@ -200,7 +201,7 @@ void Init_bson_native()
    * Returns the modified +self+.
    */
   rb_define_method(rb_byte_buffer_class, "put_int32", rb_bson_byte_buffer_put_int32, 1);
-  
+
   /*
    * call-seq:
    *   buffer.put_int64(fixnum) -> ByteBuffer
@@ -212,7 +213,7 @@ void Init_bson_native()
    * Returns the modified +self+.
    */
   rb_define_method(rb_byte_buffer_class, "put_int64", rb_bson_byte_buffer_put_int64, 1);
-  
+
   /*
    * call-seq:
    *   buffer.put_double(double) -> ByteBuffer
@@ -222,7 +223,7 @@ void Init_bson_native()
    * Returns the modified +self+.
    */
   rb_define_method(rb_byte_buffer_class, "put_double", rb_bson_byte_buffer_put_double, 1);
-  
+
   /*
    * call-seq:
    *   buffer.put_decimal128(low_64bit, high_64bit) -> ByteBuffer
@@ -235,7 +236,7 @@ void Init_bson_native()
    * Returns the modified +self+.
    */
   rb_define_method(rb_byte_buffer_class, "put_decimal128", rb_bson_byte_buffer_put_decimal128, 2);
-  
+
   /*
    * call-seq:
    *   buffer.put_hash(hash) -> ByteBuffer
@@ -245,7 +246,7 @@ void Init_bson_native()
    * Returns the modified +self+.
    */
   rb_define_method(rb_byte_buffer_class, "put_hash", rb_bson_byte_buffer_put_hash, 2);
-  
+
   /*
    * call-seq:
    *   buffer.put_array(array) -> ByteBuffer
@@ -255,7 +256,7 @@ void Init_bson_native()
    * Returns the modified +self+.
    */
   rb_define_method(rb_byte_buffer_class, "put_array", rb_bson_byte_buffer_put_array, 2);
-  
+
   /*
    * call-seq:
    *   buffer.replace_int32(position, fixnum) -> ByteBuffer
@@ -272,7 +273,7 @@ void Init_bson_native()
    * Returns the modified +self+.
    */
   rb_define_method(rb_byte_buffer_class, "replace_int32", rb_bson_byte_buffer_replace_int32, 2);
-  
+
   /*
    * call-seq:
    *   buffer.rewind! -> ByteBuffer
@@ -284,7 +285,7 @@ void Init_bson_native()
    * Returns the modified +self+.
    */
   rb_define_method(rb_byte_buffer_class, "rewind!", rb_bson_byte_buffer_rewind, 0);
-  
+
   /*
    * call-seq:
    *   buffer.to_s -> String
@@ -296,7 +297,7 @@ void Init_bson_native()
    * the buffer itself.
    */
   rb_define_method(rb_byte_buffer_class, "to_s", rb_bson_byte_buffer_to_s, 0);
-  
+
   rb_define_method(rb_bson_object_id_generator_class, "next_object_id", rb_bson_object_id_generator_next, -1);
 
   // Get the object id machine id and hash it.
@@ -309,4 +310,5 @@ void Init_bson_native()
   rb_bson_object_id_counter = FIX2INT(rb_funcall(rb_mKernel, rb_intern("rand"), 1, INT2FIX(0x1000000)));
 
   rb_bson_registry = rb_const_get(rb_bson_module, rb_intern("Registry"));
+  rb_gc_register_mark_object(rb_bson_registry);
 }

data/ext/bson/read.c

@@ -1,5 +1,5 @@
 /*
- * Copyright (C) 2009-2019 MongoDB, Inc.
+ * Copyright (C) 2009-2020 MongoDB Inc.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -17,22 +17,28 @@
 #include "bson-native.h"
 #include <ruby/encoding.h>
 
-static void pvt_validate_length(byte_buffer_t *b);
+static void pvt_raise_decode_error(volatile VALUE msg);
+static int32_t pvt_validate_length(byte_buffer_t *b);
 static uint8_t pvt_get_type_byte(byte_buffer_t *b);
 static VALUE pvt_get_int32(byte_buffer_t *b);
 static VALUE pvt_get_int64(byte_buffer_t *b, int argc, VALUE *argv);
 static VALUE pvt_get_double(byte_buffer_t *b);
-static VALUE pvt_get_string(byte_buffer_t *b);
+static VALUE pvt_get_string(byte_buffer_t *b, const char *data_type);
 static VALUE pvt_get_symbol(byte_buffer_t *b, VALUE rb_buffer, int argc, VALUE *argv);
 static VALUE pvt_get_boolean(byte_buffer_t *b);
 static VALUE pvt_read_field(byte_buffer_t *b, VALUE rb_buffer, uint8_t type, int argc, VALUE *argv);
 static void pvt_skip_cstring(byte_buffer_t *b);
 
+void pvt_raise_decode_error(volatile VALUE msg) {
+  VALUE klass = pvt_const_get_3("BSON", "Error", "BSONDecodeError");
+  rb_exc_raise(rb_exc_new_str(klass, msg));
+}
+
 /**
  * validate the buffer contains the amount of bytes the array / hash claimns
  * and that it is null terminated
  */
-void pvt_validate_length(byte_buffer_t *b)
+int32_t pvt_validate_length(byte_buffer_t *b)
 {
   int32_t length;
   
@@ -53,6 +59,8 @@ void pvt_validate_length(byte_buffer_t *b)
   else{
     rb_raise(rb_eRangeError, "Buffer contained invalid length %d at %zu", length, b->read_position);
   }
+  
+  return length;
 }
 
 /**
@@ -64,7 +72,7 @@ VALUE pvt_read_field(byte_buffer_t *b, VALUE rb_buffer, uint8_t type, int argc,
     case BSON_TYPE_INT32: return pvt_get_int32(b);
     case BSON_TYPE_INT64: return pvt_get_int64(b, argc, argv);
     case BSON_TYPE_DOUBLE: return pvt_get_double(b);
-    case BSON_TYPE_STRING: return pvt_get_string(b);
+    case BSON_TYPE_STRING: return pvt_get_string(b, "String");
     case BSON_TYPE_SYMBOL: return pvt_get_symbol(b, rb_buffer, argc, argv);
     case BSON_TYPE_ARRAY: return rb_bson_byte_buffer_get_array(argc, argv, rb_buffer);
     case BSON_TYPE_DOCUMENT: return rb_bson_byte_buffer_get_hash(argc, argv, rb_buffer);
@@ -119,9 +127,20 @@ VALUE rb_bson_byte_buffer_get_bytes(VALUE self, VALUE i)
 }
 
 VALUE pvt_get_boolean(byte_buffer_t *b){
-  VALUE result = Qnil;
+  VALUE result;
+  char byte_value;
   ENSURE_BSON_READ(b, 1);
-  result = *READ_PTR(b) == 1 ? Qtrue: Qfalse;
+  byte_value = *READ_PTR(b);
+  switch (byte_value) {
+    case 1:
+      result = Qtrue;
+      break;
+    case 0:
+      result = Qfalse;
+      break;
+    default:
+      pvt_raise_decode_error(rb_sprintf("Invalid boolean byte value: %d", (int) byte_value));
+  }
   b->read_position += 1;
   return result;
 }
@@ -134,22 +153,35 @@ VALUE rb_bson_byte_buffer_get_string(VALUE self)
   byte_buffer_t *b;
 
   TypedData_Get_Struct(self, byte_buffer_t, &rb_byte_buffer_data_type, b);
-  return pvt_get_string(b);
+  return pvt_get_string(b, "String");
 }
 
-VALUE pvt_get_string(byte_buffer_t *b)
+VALUE pvt_get_string(byte_buffer_t *b, const char *data_type)
 {
-  int32_t length;
   int32_t length_le;
+  int32_t length;
+  char *str_ptr;
   VALUE string;
+  unsigned char last_byte;
 
   ENSURE_BSON_READ(b, 4);
-  memcpy(&length, READ_PTR(b), 4);
-  length_le = BSON_UINT32_FROM_LE(length);
-  b->read_position += 4;
-  ENSURE_BSON_READ(b, length_le);
-  string = rb_enc_str_new(READ_PTR(b), length_le - 1, rb_utf8_encoding());
-  b->read_position += length_le;
+  memcpy(&length_le, READ_PTR(b), 4);
+  length = BSON_UINT32_FROM_LE(length_le);
+  if (length < 0) {
+    pvt_raise_decode_error(rb_sprintf("String length is negative: %d", length));
+  }
+  if (length == 0) {
+    pvt_raise_decode_error(rb_str_new_cstr("String length is zero but string must be null-terminated"));
+  }
+  ENSURE_BSON_READ(b, 4 + length);
+  str_ptr = READ_PTR(b) + 4;
+  last_byte = *(READ_PTR(b) + 4 + length_le - 1);
+  if (last_byte != 0) {
+    pvt_raise_decode_error(rb_sprintf("Last byte of the string is not null: 0x%x", (int) last_byte));
+  }
+  rb_bson_utf8_validate(str_ptr, length - 1, true, data_type);
+  string = rb_enc_str_new(str_ptr, length - 1, rb_utf8_encoding());
+  b->read_position += 4 + length_le;
   return string;
 }
 
@@ -166,7 +198,7 @@ VALUE pvt_get_symbol(byte_buffer_t *b, VALUE rb_buffer, int argc, VALUE *argv)
   VALUE value, klass;
 
   if (pvt_get_mode_option(argc, argv) == BSON_MODE_BSON) {
-    value = pvt_get_string(b);
+    value = pvt_get_string(b, "Symbol");
     klass = pvt_const_get_3("BSON", "Symbol", "Raw");
     value = rb_funcall(klass, rb_intern("new"), 1, value);
   } else {
@@ -306,10 +338,13 @@ VALUE rb_bson_byte_buffer_get_hash(int argc, VALUE *argv, VALUE self){
   byte_buffer_t *b = NULL;
   uint8_t type;
   VALUE cDocument = pvt_const_get_2("BSON", "Document");
+  int32_t length;
+  char *start_ptr;
 
   TypedData_Get_Struct(self, byte_buffer_t, &rb_byte_buffer_data_type, b);
 
-  pvt_validate_length(b);
+  start_ptr = READ_PTR(b);
+  length = pvt_validate_length(b);
 
   doc = rb_funcall(cDocument, rb_intern("allocate"), 0);
 
@@ -318,6 +353,11 @@ VALUE rb_bson_byte_buffer_get_hash(int argc, VALUE *argv, VALUE self){
     rb_hash_aset(doc, field, pvt_read_field(b, self, type, argc, argv));
     RB_GC_GUARD(field);
   }
+  
+  if (READ_PTR(b) - start_ptr != length) {
+    pvt_raise_decode_error(rb_sprintf("Expected to read %d bytes for the hash but read %ld bytes", length, READ_PTR(b) - start_ptr));
+  }
+  
   return doc;
 }
 
@@ -325,10 +365,13 @@ VALUE rb_bson_byte_buffer_get_array(int argc, VALUE *argv, VALUE self){
   byte_buffer_t *b;
   VALUE array = Qnil;
   uint8_t type;
+  int32_t length;
+  char *start_ptr;
 
   TypedData_Get_Struct(self, byte_buffer_t, &rb_byte_buffer_data_type, b);
 
-  pvt_validate_length(b);
+  start_ptr = READ_PTR(b);
+  length = pvt_validate_length(b);
 
   array = rb_ary_new();
   while((type = pvt_get_type_byte(b)) != 0){
@@ -336,5 +379,10 @@ VALUE rb_bson_byte_buffer_get_array(int argc, VALUE *argv, VALUE self){
     rb_ary_push(array,  pvt_read_field(b, self, type, argc, argv));
   }
   RB_GC_GUARD(array);
+  
+  if (READ_PTR(b) - start_ptr != length) {
+    pvt_raise_decode_error(rb_sprintf("Expected to read %d bytes for the hash but read %ld bytes", length, READ_PTR(b) - start_ptr));
+  }
+  
   return array;
 }