So, I need to learn not to count my chickens before they hatch.
More specifically, I need to learn not to expect language features I only have a vague understanding of to behave the way I'm expecting them to. Unfortunately, I waited until this week to start doing real research on all of the thought work that had left me with a tentative confidence that this could actually work without me completely killing myself. Now I know two things that I should have known months ago:
1. The . (dot) operator is *can't* be overloaded in C++.
2. Function pointers are messier - and more particular - than I had thought them to be.
So, my intent to make concurrent code natively look as much like normal code as possible is completely screwed, and my initial thoughts around this (particularly, writing a psuedo-parser to generate functioning C++ code from files following conventions I would determine) is making me think that I may as well just change the stated purpose of the project to something even more rooted in straight-up language design.
I wonder how well ANTLR deals with concurrency...
first, though, I suppose I should do some work with pthreads and try and see how other systems handle asynchronous returns...
Wednesday, December 24, 2008
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment