I just started learning D today and I really need to read and write data like this:
byte[] bytes = ...;
ByteBuffer buf = new ByteBuffer(bytes);
int a = buf.getInt();
byte b = buf.getByte();
short s = buf.getShort();
buf.putInt(200000);
Is there anything built in to D that can achieve this or must I make it myself?
I'd suggest looking at std.bitmanip's read, peek, write and append functions. e.g.
ubyte[] buffer = [1, 5, 22, 9, 44, 255, 8];
auto range = buffer; // if you don't want to mutate the original
assert(range.read!ushort() == 261);
assert(range == [22, 9, 44, 255, 8]);
assert(range.read!uint() == 369_700_095);
assert(range == [8]);
assert(range.read!ubyte() == 8);
assert(range.empty);
assert(buffer == [1, 5, 22, 9, 44, 255, 8]);
There is no buffer type - rather they're free functions which operate on ranges of ubyte (which includes ubyte[]) - so they may not work exactly like what you're looking for, but they are designed for the case where you need to extract integral values from an array or some other kind of range of bytes. And if you really want some kind of separate buffer type, then you should be able to create one fairly easily which uses them internally.
The old std.stream module should do the trick: http://dlang.org/phobos/std_stream.html#MemoryStream
MemoryStream buf = new MemoryStream(bytes);
// need to declare the variable ahead of time
int a;
// then read knows what to get due to the declared type
buf.read(a);
byte b;
buf.read(b);
but.write(200000); // that is an int literal so it knows it is int
buf.write(cast(ubyte) 255); // you can also explicitly cast
As the warning on the page says though, the Phobos maintainers don't like this module and want to kill it... but they've been saying that for years, I'd just go ahead and use it. Or make a private copy of the stream.d source if you like.
