Logged In: YES 
user_id=34209

Checked this in before 2.5 alpha2, after more contemplation.
Undid the circumventing of the test_pty problems, too.
Decided to explicitly set the terminating NUL after all,
instead of relying on the given buffer to have one (even
though it really should), and terminated the array in the
writable-buffer case, too (it may not matter, but it
certainly doesn't hurt, since we didn't end up copying any
data there in any case.) Thanks for figuring it out, Quentin.