bson 4.7.0 → 4.9.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: ac48ca5a022cc024d0ced56d11165c4a54d01610c1e805634510ad52db11cd7a
-  data.tar.gz: 8440d253b13a60d3af1f3e9b2179e8c376ecae89a3ea4a6cb7a3263f32d14e06
+  metadata.gz: 05c17fcaee2ceaefe97d801594460178acaf6a2f82c152ce33db819f01d624fa
+  data.tar.gz: 3343dfda6f4038fa261d077ddf860dd43ff001d2fbea7189110b6fa500fb2558
 SHA512:
-  metadata.gz: 8ffde24e9e88fd73f5e1a847c3494449ed1973348fd8407846f1a18a7671eff32ebc5df921a33fbbe0e978ed014b1bd49f93d32fb6ca213ef4bd287a2991439d
-  data.tar.gz: 1356bf7ffc9bdf12f96aa5efd32e5562eeabe1e13191354f2658f826c215833b3af30d65367c6316c955db98675ad8ee5c79ca0ec06793456fe2effd9a5700d2
+  metadata.gz: 7756cee7923f3ca584e6a1bfa3666f70d2f14804b3ddb913f0695be3495a061a632803a90d1cd87a63c95d5b9222c02b18fe475a38e2a8a680914819f90f4471
+  data.tar.gz: 88bf0b45ecbe2aa47c955ca5281568cbe6d48a4c144a5b49680f6f7d8684fa61524dcc80474118c12e9df13b6bfc4ffca1a0d0e55679c0325dcdfd48543aab96

data/README.md

@@ -1,4 +1,8 @@
-BSON [![Build Status](https://secure.travis-ci.org/mongodb/bson-ruby.png?branch=master&.png)](http://travis-ci.org/mongodb/bson-ruby)  [![Code Climate](https://codeclimate.com/github/mongodb/bson-ruby.png)](https://codeclimate.com/github/mongodb/bson-ruby) [![Coverage Status](https://coveralls.io/repos/mongodb/bson-ruby/badge.png?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 [![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)
 ====
 
 An implementation of the BSON specification in Ruby.
@@ -34,7 +38,7 @@ 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.

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.
@@ -16,6 +16,7 @@
 
 #include <ruby.h>
 #include <stdbool.h>
+#include <stdint.h>
 #include <unistd.h>
 #include <time.h>
 #include "bson-endian.h"
@@ -32,13 +33,15 @@ rb_bson_utf8_validate (const char *utf8, /* IN */
 #define HOST_NAME_HASH_MAX 256
 #endif
 
+/* See the type list in http://bsonspec.org/spec.html. */
 #define BSON_TYPE_DOUBLE        1
 #define BSON_TYPE_STRING        2
 #define BSON_TYPE_DOCUMENT      3
 #define BSON_TYPE_ARRAY         4
 #define BSON_TYPE_BOOLEAN       8
-#define BSON_TYPE_INT32         16
-#define BSON_TYPE_INT64         18
+#define BSON_TYPE_SYMBOL        0x0E
+#define BSON_TYPE_INT32         0x10
+#define BSON_TYPE_INT64         0x12
 
 typedef struct {
   size_t size;
@@ -75,8 +78,8 @@ VALUE rb_bson_byte_buffer_get_double(VALUE self);
 VALUE rb_bson_byte_buffer_get_int32(VALUE self);
 VALUE rb_bson_byte_buffer_get_int64(VALUE self);
 VALUE rb_bson_byte_buffer_get_string(VALUE self);
-VALUE rb_bson_byte_buffer_get_hash(VALUE self);
-VALUE rb_bson_byte_buffer_get_array(VALUE self);
+VALUE rb_bson_byte_buffer_get_hash(int argc, VALUE *argv, VALUE self);
+VALUE rb_bson_byte_buffer_get_array(int argc, VALUE *argv, VALUE self);
 VALUE rb_bson_byte_buffer_put_byte(VALUE self, VALUE byte);
 VALUE rb_bson_byte_buffer_put_bytes(VALUE self, VALUE bytes);
 VALUE rb_bson_byte_buffer_put_cstring(VALUE self, VALUE string);
@@ -100,6 +103,14 @@ void rb_bson_byte_buffer_free(void *ptr);
 void rb_bson_expand_buffer(byte_buffer_t* buffer_ptr, size_t length);
 void rb_bson_generate_machine_id(VALUE rb_md5_class, char *rb_bson_machine_id);
 
+VALUE pvt_const_get_2(const char *c1, const char *c2);
+VALUE pvt_const_get_3(const char *c1, const char *c2, const char *c3);
+
+#define BSON_MODE_DEFAULT       0
+#define BSON_MODE_BSON          1
+
+int pvt_get_mode_option(int argc, VALUE *argv);
+
 /**
  * The counter for incrementing object ids.
  */

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.
@@ -15,6 +15,7 @@
  */
 
 #include <string.h>
+#include <stdint.h>
 #include "bson-endian.h"
 
 /*

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.
@@ -70,8 +70,31 @@ void Init_bson_native()
   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);
-  rb_define_method(rb_byte_buffer_class, "get_hash", rb_bson_byte_buffer_get_hash, 0);
-  rb_define_method(rb_byte_buffer_class, "get_array", rb_bson_byte_buffer_get_array, 0);
+  
+  /*
+   * call-seq:
+   *   buffer.get_hash(**options) -> Hash
+   *
+   * Reads a document from the byte buffer and returns it as a BSON::Document.
+   *
+   * @option options [ nil | :bson ] :mode Decoding mode to use.
+   *
+   * @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
+   *
+   * Reads an array from the byte buffer..
+   *
+   * @option options [ nil | :bson ] :mode Decoding mode to use.
+   *
+   * @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);

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.
@@ -17,21 +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);
+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);
+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;
   
@@ -52,19 +59,23 @@ 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;
 }
 
 /**
  * Read a single field from a hash or array
  */
-VALUE pvt_read_field(byte_buffer_t *b, VALUE rb_buffer, uint8_t type){
+VALUE pvt_read_field(byte_buffer_t *b, VALUE rb_buffer, uint8_t type, int argc, VALUE *argv)
+{
   switch(type) {
     case BSON_TYPE_INT32: return pvt_get_int32(b);
-    case BSON_TYPE_INT64: return pvt_get_int64(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_ARRAY: return rb_bson_byte_buffer_get_array(rb_buffer);
-    case BSON_TYPE_DOCUMENT: return rb_bson_byte_buffer_get_hash(rb_buffer);
+    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);
     case BSON_TYPE_BOOLEAN: return pvt_get_boolean(b);
     default:
     {
@@ -116,9 +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;
 }
@@ -131,25 +153,63 @@ 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;
 }
 
+/**
+ * Reads a UTF-8 string out of the byte buffer. If the argc/argv arguments
+ * have a :mode option with the value of :bson, wraps the string in a
+ * BSON::Symbol::Raw. Otherwise consults the BSON registry to determine
+ * which class to instantiate (String in bson-ruby, overridden to Symbol by
+ * the Ruby driver). Returns either a BSON::Symbol::Raw, Symbol or String
+ * value.
+ */
+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, "Symbol");
+    klass = pvt_const_get_3("BSON", "Symbol", "Raw");
+    value = rb_funcall(klass, rb_intern("new"), 1, value);
+  } else {
+    klass = rb_funcall(rb_bson_registry, rb_intern("get"), 1, INT2FIX(BSON_TYPE_SYMBOL));
+    value = rb_funcall(klass, rb_intern("from_bson"), 1, rb_buffer);
+  }
+  
+  RB_GC_GUARD(klass);
+  return value;
+}
+
 /**
  * Get a cstring from the buffer.
  */
@@ -206,17 +266,35 @@ VALUE rb_bson_byte_buffer_get_int64(VALUE self)
 {
   byte_buffer_t *b;
   TypedData_Get_Struct(self, byte_buffer_t, &rb_byte_buffer_data_type, b);
-  return pvt_get_int64(b);
+  return pvt_get_int64(b, 0, NULL);
 }
 
-VALUE pvt_get_int64(byte_buffer_t *b)
+/**
+ * Reads a 64-bit integer out of the byte buffer into a Ruby Integer instance.
+ * If the argc/argv arguments have a :mode option with the value of :bson,
+ * wraps the integer in a BSON::Int64. Returns either the Integer or the
+ * BSON::Int64 instance.
+ */
+VALUE pvt_get_int64(byte_buffer_t *b, int argc, VALUE *argv)
 {
   int64_t i64;
+  VALUE num;
 
   ENSURE_BSON_READ(b, 8);
   memcpy(&i64, READ_PTR(b), 8);
   b->read_position += 8;
-  return LL2NUM(BSON_UINT64_FROM_LE(i64));
+  num = LL2NUM(BSON_UINT64_FROM_LE(i64));
+  
+  if (pvt_get_mode_option(argc, argv) == BSON_MODE_BSON) {
+    VALUE klass = rb_funcall(rb_bson_registry,rb_intern("get"),1, INT2FIX(BSON_TYPE_INT64));
+    VALUE value = rb_funcall(klass, rb_intern("new"), 1, num);
+    RB_GC_GUARD(klass);
+    return value;
+  } else {
+    return num;
+  }
+  
+  RB_GC_GUARD(num);
 }
 
 /**
@@ -255,40 +333,56 @@ VALUE rb_bson_byte_buffer_get_decimal128_bytes(VALUE self)
   return bytes;
 }
 
-VALUE rb_bson_byte_buffer_get_hash(VALUE self){
+VALUE rb_bson_byte_buffer_get_hash(int argc, VALUE *argv, VALUE self){
   VALUE doc = Qnil;
   byte_buffer_t *b = NULL;
   uint8_t type;
-  VALUE cDocument = rb_const_get(rb_const_get(rb_cObject, rb_intern("BSON")), rb_intern("Document"));
+  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);
+  doc = rb_funcall(cDocument, rb_intern("allocate"), 0);
 
   while((type = pvt_get_type_byte(b)) != 0){
     VALUE field = rb_bson_byte_buffer_get_cstring(self);
+    rb_hash_aset(doc, field, pvt_read_field(b, self, type, argc, argv));
     RB_GC_GUARD(field);
-    rb_hash_aset(doc, field, pvt_read_field(b, self, type));
   }
+  
+  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;
 }
 
-VALUE rb_bson_byte_buffer_get_array(VALUE self){
+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){
     pvt_skip_cstring(b);
-    rb_ary_push(array,  pvt_read_field(b, self, type));
+    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;
 }

data/ext/bson/util.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.
@@ -53,3 +53,43 @@ VALUE rb_bson_object_id_generator_next(int argc, VALUE* args, VALUE self)
   rb_bson_object_id_counter++;
   return rb_str_new(bytes, 12);
 }
+
+/**
+ * Returns a Ruby constant nested one level, e.g. BSON::Document.
+ */
+VALUE pvt_const_get_2(const char *c1, const char *c2) {
+  return rb_const_get(rb_const_get(rb_cObject, rb_intern(c1)), rb_intern(c2));
+}
+
+/**
+ * Returns a Ruby constant nested two levels, e.g. BSON::Regexp::Raw.
+ */
+VALUE pvt_const_get_3(const char *c1, const char *c2, const char *c3) {
+  return rb_const_get(pvt_const_get_2(c1, c2), rb_intern(c3));
+}
+
+/**
+ * Returns the value of the :mode option, or the default if the option is not
+ * specified. Raises ArgumentError if the value is not one of nil or :bson.
+ * A future version of bson-ruby is expected to also support :ruby and :ruby!
+ * values. Returns one of the BSON_MODE_* values.
+ */
+int pvt_get_mode_option(int argc, VALUE *argv) {
+  VALUE opts;
+  VALUE mode;
+  
+  rb_scan_args(argc, argv, ":", &opts);
+  if (NIL_P(opts)) {
+    return BSON_MODE_DEFAULT;
+  } else {
+    mode = rb_hash_lookup(opts, ID2SYM(rb_intern("mode")));
+    if (mode == Qnil) {
+      return BSON_MODE_DEFAULT;
+    } else if (mode == ID2SYM(rb_intern("bson"))) {
+      return BSON_MODE_BSON;
+    } else {
+      rb_raise(rb_eArgError, "Invalid value for :mode option: %s",
+        RSTRING_PTR(rb_funcall(mode, rb_intern("inspect"), 0)));
+    }
+  }
+}