Implement PortReader and ChanWriter

[rapha]

Hi, first pull request here so let me know if I've missed any of the procedure.

[huonw]

Indentation for the Rust code base is 4 spaces.

[rapha]

Sorry, not sure what the best way is to reply to inline comments.

I've addressed the issues you raised and squashed it back into a single commit.

[rapha]

Oops. Just realized that buffering this way will typically cause the buffer to grow endlessly. I guess a chunked buffer that can release chunks as they are read would be better.

[alexcrichton]

Awesome work, thanks for doing this!

[brson]

Neat! I forgot these types were around and unimplemented.

[rapha]

So, I've updated the code with more doc, changed the way it buffers to reduce copying, and fixed eof.

Sorry, I realise I shouldn't have been continually squashing my changes back into a single commit, as this breaks the whole pull request mechansim.

Please take another look.

[huonw]

Does this work on a recent master? I believe with the move to use proc()'s, spawn_with is now unnecessary (and is removed), i.e.

task::spawn(proc() {
   let mut reader = PortReader::new(port);
   // etc.
})

should work fine.

[alexcrichton]

This should also invert the sender/receiver. Right now failures in spawned tasks won't trigger test failures, so the asserts should happen in the test task (unless you have a channel waiting for the test task to exit, in which case a failure would close the channel causing a test failure).

[bill-myers]

How about generalizing this to work on an Iterator instead of a Port, and implementing Iterator for Port?

Then one could use it for a [[u8]] as well, for instance.

[rapha]

@bill-myers That's a good idea. However, when I tried to

impl<P: GenericPort<~[u8]>> iter::Iterator<~[u8]> for P

rustc complained about a lot of conflicting implementations. This works

impl iter::Iterator<~[u8]> for Port<~[u8]>

but doesn't seem like the right solution.

I might leave that for a separate pull request.

[alexcrichton]

I think this should also factor in whether self.buf is None vs Some and where pos is relative to the array inside Some. Could you add a test for this as well?

[rapha]

self.closed can only be true if self.buf is None. That is currently tested in test_port_reader.

[alexcrichton]

good point!

[alexcrichton]

I think all that's left is a squash of commits, then r+ from me.

[alexcrichton]

Nice work! One small fix to eof and r+ by me. Also, would you mind rebasing these commits into a single commit? Thanks again!

[rapha]

Squashed.