|
|
 | | Re: Design by Contract (long post)
|
| (...) I agree it seems to be the wrong way round. That's the standard implementation. NDEBUG seems to stand for "No debug". One advantage of making it negative, is that you get the assert()s if you don't do anything (ie dont define NDEBUG). BTW, (...) (22 years ago, 8-Jan-04, to lugnet.robotics.rcx, FTX)
| | | | | Re: Design by Contract (long post)
|
| (...) D'oH! Thanks, Tim. I was very careful composing the original post, but added the example in as an after thought - Well spotted. Iain. (22 years ago, 8-Jan-04, to lugnet.robotics.rcx, FTX)
| | | | | RE: NQC's future
|
| (...) Ha Ha, very funny :-) Seriously, Philippe, why not give pbForth a try? I know that NQC works great on both the Spybots and the RCX, and even Cybermaster. But if you're looking for speed, higher precision, lots of variables and generally more (...) (22 years ago, 8-Jan-04, to lugnet.robotics.rcx.nqc, lugnet.robotics.spybotics, lugnet.robotics.rcx.pbforth)
| | | | | Re: NQC's future
|
| (...) Something I'd like to see implemented in NQC is a mixed-mode multiply/divide, similar to the */ operator in Forth, with an intermediary 32bits product. That would greatly ease precision calculations without requiring major structural (...) (22 years ago, 8-Jan-04, to lugnet.robotics.rcx.nqc, lugnet.robotics.spybotics)
| | | | | RE: IR Tower and Hyperthreading
|
| <SNIP> (...) It should be illegal to use "Linux" and "straight forward" in the same sentence. Notice, everything after that line uses words like "scattered", "most helpful", "other flavors", "to be added to the kernel [in upcoming release]", and (...) (22 years ago, 8-Jan-04, to lugnet.robotics)
|
|
