Add binary COPY encoders and decoders and binary timestamp encoders

README.md

@@ -125,8 +125,9 @@ to convert single values to/from their string representation.
 The following PostgreSQL column types are supported by ruby-pg (TE = Text Encoder, TD = Text Decoder, BE = Binary Encoder, BD = Binary Decoder):
 
 * Integer: [TE](rdoc-ref:PG::TextEncoder::Integer), [TD](rdoc-ref:PG::TextDecoder::Integer), [BD](rdoc-ref:PG::BinaryDecoder::Integer) 💡 No links? Switch to [here](https://deveiate.org/code/pg/README_md.html#label-Type+Casts) 💡
-  * BE: [Int2](rdoc-ref:PG::BinaryEncoder::Int2), [Int4](rdoc-ref:PG::BinaryEncoder::Int4), [Int8](rdoc-ref:PG::BinaryEncoder::Int8)
+    * BE: [Int2](rdoc-ref:PG::BinaryEncoder::Int2), [Int4](rdoc-ref:PG::BinaryEncoder::Int4), [Int8](rdoc-ref:PG::BinaryEncoder::Int8)
 * Float: [TE](rdoc-ref:PG::TextEncoder::Float), [TD](rdoc-ref:PG::TextDecoder::Float), [BD](rdoc-ref:PG::BinaryDecoder::Float)
+    * BE: [Float4](rdoc-ref:PG::BinaryEncoder::Float4), [Float8](rdoc-ref:PG::BinaryEncoder::Float8)
 * Numeric: [TE](rdoc-ref:PG::TextEncoder::Numeric), [TD](rdoc-ref:PG::TextDecoder::Numeric)
 * Boolean: [TE](rdoc-ref:PG::TextEncoder::Boolean), [TD](rdoc-ref:PG::TextDecoder::Boolean), [BE](rdoc-ref:PG::BinaryEncoder::Boolean), [BD](rdoc-ref:PG::BinaryDecoder::Boolean)
 * String: [TE](rdoc-ref:PG::TextEncoder::String), [TD](rdoc-ref:PG::TextDecoder::String), [BE](rdoc-ref:PG::BinaryEncoder::String), [BD](rdoc-ref:PG::BinaryDecoder::String)
@@ -135,16 +136,17 @@ The following PostgreSQL column types are supported by ruby-pg (TE = Text Encode
 * Timestamp:
     * TE: [local](rdoc-ref:PG::TextEncoder::TimestampWithoutTimeZone), [UTC](rdoc-ref:PG::TextEncoder::TimestampUtc), [with-TZ](rdoc-ref:PG::TextEncoder::TimestampWithTimeZone)
     * TD: [local](rdoc-ref:PG::TextDecoder::TimestampLocal), [UTC](rdoc-ref:PG::TextDecoder::TimestampUtc), [UTC-to-local](rdoc-ref:PG::TextDecoder::TimestampUtcToLocal)
+    * BE: [local](rdoc-ref:PG::BinaryEncoder::TimestampLocal), [UTC](rdoc-ref:PG::BinaryEncoder::TimestampUtc)
     * BD: [local](rdoc-ref:PG::BinaryDecoder::TimestampLocal), [UTC](rdoc-ref:PG::BinaryDecoder::TimestampUtc), [UTC-to-local](rdoc-ref:PG::BinaryDecoder::TimestampUtcToLocal)
 * Date: [TE](rdoc-ref:PG::TextEncoder::Date), [TD](rdoc-ref:PG::TextDecoder::Date)
 * JSON and JSONB: [TE](rdoc-ref:PG::TextEncoder::JSON), [TD](rdoc-ref:PG::TextDecoder::JSON)
 * Inet: [TE](rdoc-ref:PG::TextEncoder::Inet), [TD](rdoc-ref:PG::TextDecoder::Inet)
 * Array: [TE](rdoc-ref:PG::TextEncoder::Array), [TD](rdoc-ref:PG::TextDecoder::Array)
 * Composite Type (also called "Row" or "Record"): [TE](rdoc-ref:PG::TextEncoder::Record), [TD](rdoc-ref:PG::TextDecoder::Record)
 
-The following text formats can also be encoded although they are not used as column type:
+The following text and binary formats can also be encoded although they are not used as column type:
 
-* COPY input and output data: [TE](rdoc-ref:PG::TextEncoder::CopyRow), [TD](rdoc-ref:PG::TextDecoder::CopyRow)
+* COPY input and output data: [TE](rdoc-ref:PG::TextEncoder::CopyRow), [TD](rdoc-ref:PG::TextDecoder::CopyRow), [BE](rdoc-ref:PG::BinaryEncoder::CopyRow), [BD](rdoc-ref:PG::BinaryDecoder::CopyRow)
 * Literal for insertion into SQL string: [TE](rdoc-ref:PG::TextEncoder::QuotedLiteral)
 * SQL-Identifier: [TE](rdoc-ref:PG::TextEncoder::Identifier), [TD](rdoc-ref:PG::TextDecoder::Identifier)
 

ext/pg.c

@@ -33,7 +33,6 @@
  *
  * - PQfreemem -- unnecessary: copied to ruby object, then freed. Ruby object's
  *                memory is freed when it is garbage collected.
- * - PQbinaryTuples -- better to use PQfformat
  * - PQprint -- not very useful
  * - PQsetdb -- not very useful
  * - PQoidStatus -- deprecated, use PQoidValue

ext/pg.h

@@ -314,7 +314,7 @@ int pg_coder_enc_to_s                                  _(( t_pg_coder*, VALUE, c
 int pg_text_enc_identifier                             _(( t_pg_coder*, VALUE, char *, VALUE *, int));
 t_pg_coder_enc_func pg_coder_enc_func                  _(( t_pg_coder* ));
 t_pg_coder_dec_func pg_coder_dec_func                  _(( t_pg_coder*, int ));
-void pg_define_coder                                   _(( const char *, void *, VALUE, VALUE ));
+VALUE pg_define_coder                                  _(( const char *, void *, VALUE, VALUE ));
 VALUE pg_obj_to_i                                      _(( VALUE ));
 VALUE pg_tmbc_allocate                                 _(( void ));
 void pg_coder_init_encoder                             _(( VALUE ));

ext/pg_binary_encoder.c

@@ -93,6 +93,131 @@ pg_bin_enc_int8(t_pg_coder *conv, VALUE value, char *out, VALUE *intermediate, i
 	return 8;
 }
 
+/*
+ * Document-class: PG::BinaryEncoder::Float4 < PG::SimpleEncoder
+ *
+ * This is the binary encoder class for the PostgreSQL +float4+ type.
+ *
+ */
+static int
+pg_bin_enc_float4(t_pg_coder *conv, VALUE value, char *out, VALUE *intermediate, int enc_idx)
+{
+	union {
+		float f;
+		int32_t i;
+	} swap4;
+
+	if(out){
+		swap4.f = NUM2DBL(*intermediate);
+		write_nbo32(swap4.i, out);
+	}else{
+		*intermediate = value;
+	}
+	return 4;
+}
+
+/*
+ * Document-class: PG::BinaryEncoder::Float8 < PG::SimpleEncoder
+ *
+ * This is the binary encoder class for the PostgreSQL +float8+ type.
+ *
+ */
+static int
+pg_bin_enc_float8(t_pg_coder *conv, VALUE value, char *out, VALUE *intermediate, int enc_idx)
+{
+	union {
+		double f;
+		int64_t i;
+	} swap8;
+
+	if(out){
+		swap8.f = NUM2DBL(*intermediate);
+		write_nbo64(swap8.i, out);
+	}else{
+		*intermediate = value;
+	}
+	return 8;
+}
+
+#define PG_INT64_MIN	(-0x7FFFFFFFFFFFFFFFL - 1)
+#define PG_INT64_MAX	0x7FFFFFFFFFFFFFFFL
+
+/*
+ * Document-class: PG::BinaryEncoder::Timestamp < PG::SimpleEncoder
+ *
+ * This is a encoder class for conversion of Ruby Time objects to PostgreSQL binary timestamps.
+ *
+ * The following flags can be used to specify timezone interpretation:
+ * * +PG::Coder::TIMESTAMP_DB_UTC+ : Send timestamp as UTC time (default)
+ * * +PG::Coder::TIMESTAMP_DB_LOCAL+ : Send timestamp as local time (slower)
+ *
+ * Example:
+ *   enco = PG::BinaryEncoder::Timestamp.new(flags: PG::Coder::TIMESTAMP_DB_UTC)
+ *   enco.encode(Time.utc(2000, 1, 1))  # => "\x00\x00\x00\x00\x00\x00\x00\x00"
+ *
+ * String values are expected to contain a binary data with a length of 8 byte.
+ *
+ */
+static int
+pg_bin_enc_timestamp(t_pg_coder *this, VALUE value, char *out, VALUE *intermediate, int enc_idx)
+{
+	if(out){
+		int64_t timestamp;
+		struct timespec ts;
+
+		/* second call -> write data to *out */
+		switch(TYPE(*intermediate)){
+			case T_STRING:
+				return pg_coder_enc_to_s(this, value, out, intermediate, enc_idx);
+			case T_TRUE:
+				write_nbo64(PG_INT64_MAX, out);
+				return 8;
+			case T_FALSE:
+				write_nbo64(PG_INT64_MIN, out);
+				return 8;
+		}
+
+		ts = rb_time_timespec(*intermediate);
+		/* PostgreSQL's timestamp is based on year 2000 and Ruby's time is based on 1970.
+			* Adjust the 30 years difference. */
+		timestamp = (ts.tv_sec - 10957L * 24L * 3600L) * 1000000 + (ts.tv_nsec / 1000);
+
+		if( this->flags & PG_CODER_TIMESTAMP_DB_LOCAL ) {
+			/* send as local time */
+			timestamp += NUM2LL(rb_funcall(*intermediate, rb_intern("utc_offset"), 0)) * 1000000;
+		}
+
+		write_nbo64(timestamp, out);
+	}else{
+		/* first call -> determine the required length */
+		if(TYPE(value) == T_STRING){
+			char *pstr = RSTRING_PTR(value);
+			if(RSTRING_LEN(value) >= 1){
+				switch(pstr[0]) {
+					case 'I':
+					case 'i':
+						*intermediate = Qtrue;
+						return 8;
+					case '-':
+						if (RSTRING_LEN(value) >= 2 && (pstr[1] == 'I' || pstr[1] == 'i')) {
+							*intermediate = Qfalse;
+							return 8;
+						}
+				}
+			}
+
+			return pg_coder_enc_to_s(this, value, out, intermediate, enc_idx);
+		}
+
+		if( this->flags & PG_CODER_TIMESTAMP_DB_LOCAL ) {
+			/* make a local time, so that utc_offset is set */
+			value = rb_funcall(value, rb_intern("getlocal"), 0);
+		}
+		*intermediate = value;
+	}
+	return 8;
+}
+
 /*
  * Document-class: PG::BinaryEncoder::FromBase64 < PG::CompositeEncoder
  *
@@ -153,10 +278,16 @@ init_pg_binary_encoder(void)
 	pg_define_coder( "Int4", pg_bin_enc_int4, rb_cPG_SimpleEncoder, rb_mPG_BinaryEncoder );
 	/* dummy = rb_define_class_under( rb_mPG_BinaryEncoder, "Int8", rb_cPG_SimpleEncoder ); */
 	pg_define_coder( "Int8", pg_bin_enc_int8, rb_cPG_SimpleEncoder, rb_mPG_BinaryEncoder );
+	/* dummy = rb_define_class_under( rb_mPG_BinaryEncoder, "Float4", rb_cPG_SimpleEncoder ); */
+	pg_define_coder( "Float4", pg_bin_enc_float4, rb_cPG_SimpleEncoder, rb_mPG_BinaryEncoder );
+	/* dummy = rb_define_class_under( rb_mPG_BinaryEncoder, "Float8", rb_cPG_SimpleEncoder ); */
+	pg_define_coder( "Float8", pg_bin_enc_float8, rb_cPG_SimpleEncoder, rb_mPG_BinaryEncoder );
 	/* dummy = rb_define_class_under( rb_mPG_BinaryEncoder, "String", rb_cPG_SimpleEncoder ); */
 	pg_define_coder( "String", pg_coder_enc_to_s, rb_cPG_SimpleEncoder, rb_mPG_BinaryEncoder );
 	/* dummy = rb_define_class_under( rb_mPG_BinaryEncoder, "Bytea", rb_cPG_SimpleEncoder ); */
 	pg_define_coder( "Bytea", pg_coder_enc_to_s, rb_cPG_SimpleEncoder, rb_mPG_BinaryEncoder );
+	/* dummy = rb_define_class_under( rb_mPG_BinaryEncoder, "Timestamp", rb_cPG_SimpleEncoder ); */
+	pg_define_coder( "Timestamp", pg_bin_enc_timestamp, rb_cPG_SimpleEncoder, rb_mPG_BinaryEncoder );
 
 	/* dummy = rb_define_class_under( rb_mPG_BinaryEncoder, "FromBase64", rb_cPG_CompositeEncoder ); */
 	pg_define_coder( "FromBase64", pg_bin_enc_from_base64, rb_cPG_CompositeEncoder, rb_mPG_BinaryEncoder );

ext/pg_coder.c

@@ -455,7 +455,7 @@ static const rb_data_type_t pg_coder_cfunc_type = {
 	RUBY_TYPED_FREE_IMMEDIATELY,
 };
 
-void
+VALUE
 pg_define_coder( const char *name, void *func, VALUE base_klass, VALUE nsp )
 {
 	VALUE cfunc_obj = TypedData_Wrap_Struct( rb_cObject, &pg_coder_cfunc_type, func );
@@ -471,6 +471,7 @@ pg_define_coder( const char *name, void *func, VALUE base_klass, VALUE nsp )
 	rb_define_const( coder_klass, "CFUNC", cfunc_obj );
 
 	RB_GC_GUARD(cfunc_obj);
+	return coder_klass;
 }