 | | Re: NQC wishlist
|
|
(...) I believe the simplest way to do it would be to have an extra preprocessing pass... it will slow down compilation a little, it's true, but maybe not that much as to matter (compilation is anyway so much faster than the downloading anyway) then (...) (25 years ago, 5-Jan-00, to lugnet.robotics.rcx.nqc)
|
| |
| | Re: NQC wishlist
|
|
(...) No, this sort of thing isn't covered in my book. There are several ways to send commands to the RCX. The first is the official SDK from TLG (spirit.ocx) which runs under Windows. The SDK contains a reasonable amount of documentation on the (...) (25 years ago, 5-Jan-00, to lugnet.robotics.rcx, lugnet.robotics.rcx.nqc)
|
| |
| | Re: NQC wishlist
|
|
(...) [Snip] (...) Okay. Is sending these commands (and which ones that are/can be "immediate") in your book? <G> While I don not have something in mind right this second, the more complex programs get the higher the possibility of someone using (...) (25 years ago, 4-Jan-00, to lugnet.robotics.rcx, lugnet.robotics.rcx.nqc)
|
| |
| | Re: NQC wishlist
|
|
(...) I'm not sure if the terminology from LEGO is "direct" or "immediate" commands, but either way the idea is that some of the bytecodes can be sent to the RCX and the RCX will execute them immediately. There's a lot of overlap between the direct (...) (25 years ago, 4-Jan-00, to lugnet.robotics.rcx, lugnet.robotics.rcx.nqc)
|
| |
| | Re: NQC wishlist
|
|
(...) I am kind of new to this, but what is meant by direct command? How would tasks downloaded in another slot be assessable to some other task to start up? --- DonC donc@cccd.edu (25 years ago, 4-Jan-00, to lugnet.robotics.rcx, lugnet.robotics.rcx.nqc)
|
| |
| | RE: NQC wishlist
|
|
(...) This might be a good idea. The pbForth thing is working out well too. Maybe a C to FORTH translator would be useful. In the other hand, making a custom bytecode interpreter means having to write the interpreter and support it on different (...) (25 years ago, 4-Jan-00, to lugnet.robotics.rcx, lugnet.robotics.rcx.nqc)
|
| |
| | Re: NQC wish
|
|
Inline functions are by their very nature surpressed. Also, most control structures and calculations that can be evaluated at compile time are eliminated. I thought about surpressing tasks and subs, but then decided not to. The problem is that tasks (...) (25 years ago, 4-Jan-00, to lugnet.robotics.rcx.nqc)
|
| |
| | Re: NQC wishlist
|
|
(...) To me, one of the important features of such a thing would be compatibility with the existing firmware. Otherwise, why not just use LegOS? (25 years ago, 4-Jan-00, to lugnet.robotics.rcx, lugnet.robotics.rcx.nqc)
|
| |
| | Re: NQC wishlist
|
|
If anyone is seriously considering this I'd ask two things... 1) Have you considered porting the p-code interpreter used by Interactive C or something like a stripped down Java bytecode interpreter instead? The idea here would be that the RCX (...) (25 years ago, 4-Jan-00, to lugnet.robotics.rcx, lugnet.robotics.rcx.nqc)
|
| |
| | Re: NQC wishlist
|
|
(...) Yeah, this was discussed way back at the beginning of the reverse-engineering effort. It's something I'd like to see too. (25 years ago, 4-Jan-00, to lugnet.robotics.rcx, lugnet.robotics.rcx.nqc)
|
| |
| | Re: NQC wishlist
|
|
I was jut wondering, has anybody tried to implement a "better RCX-code"? It must be possible to write something very similar to RCX-code in LegOS, only faster and implementing the "missing" array structures and more variables. Mayby compatible with (...) (25 years ago, 4-Jan-00, to lugnet.robotics.rcx, lugnet.robotics.rcx.nqc)
|
| |
| | NQC wish
|
|
Hello, It would be great to have NCQ doing dead code suppression before downloading it to the RCX. Any function, task,... not referenced in the program should not be compiled and sent to the RCX. It would help to save memory and download time don't (...) (25 years ago, 4-Jan-00, to lugnet.robotics.rcx.nqc)
|
| |
| | Re: NQC wishlist
|
|
(...) Implementation of ## gets pretty nasty - at least within the current pre-processor design. I'll look at the C spec again, but I'm pretty sure ## forces a re-tokenization. In the current design of NQC, tokenization happens before (...) (25 years ago, 4-Jan-00, to lugnet.robotics.rcx.nqc)
|
| |
| | Re: NQC wishlist
|
|
(...) Great! (...) Vlad, that's funny. Immediately after seeing Dave's reply, I decided to make such macros myself. But I got disturbed and couldn't do it today anymore. Since you seem to have done it already, I'll just wait to see yours :-) (...) (...) (25 years ago, 4-Jan-00, to lugnet.robotics.rcx.nqc)
|
| |
| | Re: NQC wishlist
|
|
(...) I am sure more people than me are looking forward to that! About the "smaller" integer variables, I have put together a couple of macros that simulate an array of packed "small" integers, any bit size works, but power-of-2 sizes do not waste (...) (25 years ago, 4-Jan-00, to lugnet.robotics.rcx.nqc)
|
| |
| | Re: What should be in the next version of RcxCC
|
|
(...) Thank you Mark for your great tool. Several things could improve Rcx CC. - An indentation of selected lines with Tab or Shift+Tab like in Visual C++ editor. When a lines are selected, Tab inserts tabs at the beginning of the lines, and (...) (25 years ago, 4-Jan-00, to lugnet.robotics.rcx.nqc)
|
| |
| | Re: NQC wishlist
|
|
A switch statment will be added pretty soon. The variable stuff is on hold for the moment. I'm holding off on the booleans until I work out a better strategy for variable allocation and code generation in general. Dave (...) for X-mas), (...) nested (...) (25 years ago, 3-Jan-00, to lugnet.robotics.rcx.nqc)
|
| |
| | Re: NQC wishlist
|
|
(...) Hi Dave, I'm using your wonderful tool since just a week now (bought my first RIS for X-mas), and I am already seriously running out of variables ... :-) So, I think your idea of variable types of less than 16 bits such as booles and shorts is (...) (25 years ago, 3-Jan-00, to lugnet.robotics.rcx.nqc)
|
| |
| | Re: What should be in the next version of RcxCC
|
|
(...) Oops, that's right. I completely overlooked this. Thanks for the hint. So, my previous suggestion is quite pointless. Please ignore it. Sorry. Uwe (25 years ago, 3-Jan-00, to lugnet.robotics.rcx.nqc)
|
| |
| | Re: What should be in the next version of RcxCC
|
|
You can also choose "show code/error listing" under the compile menu to get the variable assignments. Mike Uwe Denzer <Uwe.Denzer@munich.netsurf.de> wrote in message news:386F928E.984386...surf.de... (...) like to (...) up in (...) and I'd (...) (...) (25 years ago, 3-Jan-00, to lugnet.robotics.rcx.nqc)
|
