2023-08-15 00:04:44 Maybe. It ran for three more seasons, though. 2023-08-15 00:04:56 No writere's strike or anything like that - I checked. 2023-08-15 00:08:40 Near the very end of that series they threw in a subplot that involved Magnum suspecting that Higgins actually *was* Robin Master (the wealthy writer that owned the estate Magnum lived on and Higgins managed). 2023-08-15 00:08:49 The never resolved that - left the audience wondering about it. 2023-08-15 00:09:04 But if you watch the early seasons carefully you see things that would seem to make that an impossibility. 2023-08-15 00:09:24 The audience got to see at least one phone call Higgins took from Robin when no one else was around to watch. 2023-08-15 09:43:23 Hrm. Not sure I'm a fan of SmithForth's naming here. It's 64-bit, so @ fetches 8 bytes, and we have the standard C@ to fetch a single byte. 2023-08-15 09:43:28 Then there's the standard 2@ to get two cells worth of memory. 2023-08-15 09:43:30 But SmithForth has a non-standard word for fetching a doubleword (i.e. 4 bytes), 4@. 2023-08-15 09:44:22 So 2@ reads 4x the number of bytes than 4@ :P 2023-08-15 09:45:48 Well, I had to ban a Dresden community member this morning. Someone asked about where they could get the graphic novels, and this guy responds with a comment where he had put certain letters of his reply in bold capital letters, and they literally spelled out the URL of a comic book piracy site. :-( 2023-08-15 09:46:03 I don't know if he thought he was being clever, but... gee. 2023-08-15 09:46:37 The rules stipulate an automatic peramanent ban for soliciting / encouraging / aiding piracy. 2023-08-15 09:47:11 I don't expect "no one to ever do it," but come on - don't do it right there in front of the whole world. :-( 2023-08-15 09:48:09 I'd never heard of the site, but all I had to do was type his bold cap letters into my address bar and there it was. 2023-08-15 09:48:25 He even made sure to include the ".org" 2023-08-15 09:49:05 he's too kind :\ 2023-08-15 09:50:02 For his own good, yes. 2023-08-15 09:50:24 Well, technically folks aren't supposed to come back in under another account in these situations, but I'm sure it happens. 2023-08-15 09:51:01 And unless he's dumb about that too (say, talks about it in open comments), he could get away with it. So I don't feel too bad. 2023-08-15 09:51:29 IPs can be banned too, but that's a higher tier action; admin level. 2023-08-15 16:38:54 I think I'm going to try to write an assembly vm for the 32655 that's wrapped inside a C program I can run from my dev system. Have C initialize a stack and "start" the vm. Allocate an unused opcode to "return to C," dump the stack and whatever RAM might be of interest. 2023-08-15 17:20:20 vms14, veltas: No you can not forget a closure!! mk-counter?! More than just a counter Closure tick, reset, get, and set can be packaged into a (Hash?) Table! A Javacript ""Object""!? 2023-08-15 17:27:36 rendar: Clearly You need to clarify the Arity!! Currying?! So Your example would get a n-Tuple! 2-Tuple in first example and 3-Tuple in second example. But how do You _Construct_ those Tuples!? Take a look at Python ""argument""... mechanism, positional and keyword. It is convenient. Python ""function"" is merely a Syntactic Sugar for ""dict"" get and set. 2023-08-15 17:29:50 ""function"" ""call"" initializes the Table, ""dict"", and ""function"" ""body"" Operates on it. 2023-08-15 17:33:05 ""global"" ""variable""s are special case of ""function"" ""parameter""s. 2023-08-15 17:34:02 rendar: ((in short) Your answer is) ""Activation Record""?! 2023-08-15 18:42:36 KipIngram: interesting. what's the advantage of the VM in this case? 2023-08-15 18:44:44 Oh, well, eventually that will be the VM that runs my whole 32655 Forth system. When properly deployed it won't be wraped in C anymore, but will just be part of the image I burn into the thing. The purposes are 1) portability - I intend to write as much of the Forth system as possible in the vm instruction set - and 2) compactness. This vm will run six-bit opcodes that are packed 5 and occasionally six (in 2023-08-15 18:44:46 lucky cases) into each 32-bit cell. 2023-08-15 18:45:12 It's modeled on the F18A instruction set. The F18A uses 5-bit opcodes and packs 3 / sometimes 4 into each 18-bit cell. Same idea though. 2023-08-15 18:45:14 I'm interested in using a VM to squeeze a big program onto a small microcontroller. was thinking compiling to some other random architecture and using that for VM could save space. 2023-08-15 18:45:26 I need more that 32 instructions to do it how I want to do it, but 64 is enough. 2023-08-15 18:45:44 ah I see 2023-08-15 18:45:46 That occasional "lucky" extra instruction is when you happen to need an instruction that has the high bits zero. 2023-08-15 18:48:58 This plan with the C wrapper is just a path for developing and debugging the thing. 2023-08-15 18:49:48 And I imagine I can go quite a long way using it - not just get the vm running, but also develop more or less the entire system. Then just rip it out of the wrapper at the last minute to make a finished image. 2023-08-15 18:50:15 KipIngram: 32-64 instructions? 2023-08-15 18:50:52 It's a 32-bit processor. I'll have 32-bit cells and 32-bit stacks. I'll decide at some point whether to include double int instructions or not. 2023-08-15 18:51:13 I expect to use most but not all of the 64 opcodes. 2023-08-15 18:51:40 The F18A uses all 32 of its possible opcodes, but I just need a little more for some of my "additions." 2023-08-15 18:51:43 got your vm spec frozen down? 2023-08-15 18:51:52 No, not yet. 2023-08-15 18:52:13 Haven't actually gone through and listed out all my instructions yet, but I have a pretty good feel for what I want. 2023-08-15 18:52:36 The only part that's a little depressing is how to approach 16 and 8 bit RAM operations. 2023-08-15 18:52:50 I ask because I like to implement emulators for such in my fcpu32 system 2023-08-15 18:53:20 One approach is just to add more opcodes as needed for those, but that will use up almost half of my extra 32 opcodes. 2023-08-15 18:53:38 I may do it that way, though, if my other additions still fit. 2023-08-15 18:54:19 been petting the idea of moving most of the software I write ontop of fcpu32 vm + cloudabi subset of posix 2023-08-15 18:54:44 (graphics via vnc basically) 2023-08-15 19:04:06 you looking for speed ups by having more opcodes? 2023-08-15 19:43:13 Yes, some of the things I've added I want to be at the lowest level. A subset of the conditional returns, etc. 2023-08-15 20:08:48 You could have a flag to select data size like some processors 2023-08-15 20:09:09 might make it annoying to program but it would save space 2023-08-15 20:11:00 you'd need a convention for saving or restoring the mode when you call into words 2023-08-15 21:35:43 That's the other option, right. Just the one set of access words, and then three additional instructions to set the size. 2023-08-15 21:36:35 It could either stick until changed or just apply to one access. If I did it the latter way I'd only need two size selectors; "absent" would get the third (32-bit) size. 2023-08-15 21:37:00 I've been thinking all these very things and haven't made any decisions yet. 2023-08-15 21:37:39 And since my threads can't be interrupted between opcodes, that aspect wouldn't be a problem. 2023-08-15 23:38:08 So how does memory access work on Arm? Are you allowed, for example, to fetch or store a 32-bit quantity that isn't 32-bit aligned? Or a 16-bit quantity?