2021-11-11 04:25:41 <neuro_sys_  > Forgot to respond yesterday, POSTPONE and ['] are clear now.
2021-11-11 04:26:02 <neuro_sys_  > I was later looking into how DOES> is implemented. Forth Programmer's Manual has been more helpful than reading the 94 standard.
2021-11-11 04:28:11 <f-a         > who is the author
2021-11-11 04:33:50 <neuro_sys_  > Ah, I meant to say *Handbook, not Manual.
2021-11-11 08:06:06 <veltas      > neuro_sys_: The standards are not generally helpful for understanding how FORTH is implemented
2021-11-11 08:06:33 <veltas      > The Moving FORTH articles have info on DOES>
2021-11-11 08:45:16 <neuro_sys_  > Oh right, I missed the next articles in Moving FORTH
2021-11-11 10:58:52 <neuro_sys_  > https://www.cs.utexas.edu/users/EWD/transcriptions/EWD00xx/EWD28.html
2021-11-11 12:23:51 <joe9        > anyone aware of any good articles on exceptions in forth? something catering to the starting forth crowd (to which I belong).
2021-11-11 13:13:08 <veltas      > joe9: What do you want to know about exceptions?
2021-11-11 13:33:40 <veltas      > joe9: Is this helpful? https://ideone.com/Qezi20
2021-11-11 13:36:46 <veltas      > The syntax highlighting worked fine in their editor, but displays bad when it's submitted...
2021-11-11 13:39:52 <imode       > what's `throw` do.
2021-11-11 13:41:46 <imode       > does that rely on a separate exception stack
2021-11-11 13:41:53 <imode       > it looks like it..
2021-11-11 13:47:32 <veltas      > https://forth-standard.org/standard/exception
2021-11-11 13:47:34 <veltas      > Yes
2021-11-11 13:47:49 <imode       > seems like we just keep adding stacks.
2021-11-11 13:48:13 <veltas      > I think the main use is a cleaner rdrop exit when you've got many places that want to roll back to one calling word
2021-11-11 13:48:24 <veltas      > And as a way of implementing a more generic abort
2021-11-11 13:48:45 <veltas      > This is one stack I don't mind adding, and I'd only expect it to be 4 or so entries deep
2021-11-11 13:48:55 <imode       > that's fair.
2021-11-11 13:49:14 <veltas      > Well, maybe deeper than that, but it doesn't need to be very deep
2021-11-11 13:49:33 <imode       > I suspect none of the stacks should be very deep.
2021-11-11 13:49:44 <veltas      > For example, I'm currently writing some code that does some parsing, and I am using exceptions to handle a parse failure
2021-11-11 13:52:50 <veltas      > So anyway, CATCH will take an xt, execute it, if THROW is called during that then the parameter stack and return stack are reset and execution resumes following CATCH. CATCH always leaves a number on parameter stack, it's 0 if no exception and the exception code otherwise
2021-11-11 13:53:00 <veltas      > THROW takes one number, if it's non-zero an exception is thrown
2021-11-11 13:53:18 <veltas      > ABORT is redefined to essentially do -1 THROW
2021-11-11 13:53:46 <veltas      > ABORT" similarly does -2 THROW
2021-11-11 13:54:40 <veltas      > That's all there is to it, and a stack is needed for the exception handling to work, the "exception stack", which will probably be a small buffer somewhere with entries capable of restoring IP/SP/RP
2021-11-11 13:55:56 <imode       > that feels like half an exception system in contrast with C/C++.
2021-11-11 13:57:46 <joe9        > veltas, imode, thanks.
2021-11-11 13:58:40 <veltas      > imode: It's really just setjmp
2021-11-11 13:59:06 <veltas      > C++ exceptions have to worry about RAII yeah
2021-11-11 13:59:48 <imode       > yeah
2021-11-11 14:00:08 <veltas      > I think that my exception implementation was going to put the info on the return stack, and then just have one pointer somewhere to the current CATCH data (or 0 if no CATCH)
2021-11-11 14:00:11 <eris[m]     > you can make your own system
2021-11-11 14:00:26 <veltas      > Because then you don't need a new stack for it, it can fit in return stack
2021-11-11 14:04:52 <veltas      > Standard says the exception stack may be the return stack
2021-11-11 14:05:58 <imode       > curious.
2021-11-11 14:07:44 <veltas      > Probably that's how everyone does it, so really there is no exception stack
2021-11-11 14:08:02 <veltas      > Because thinking about it, there's no need. You just need a pointer to the last exception frame on the return stack
2021-11-11 14:08:04 <imode       > that makes sense..
2021-11-11 14:08:21 <imode       > yeah. you're only ever using the top half when the exception system is triggered.
2021-11-11 14:08:23 <veltas      > That's how C does it too
2021-11-11 14:20:27 <veltas      > I'm going to try writing something for simple linear regression in FORTH since someone asked on Facebook
2021-11-11 14:20:33 <veltas      > If anyone else is interested in giving it a shot
2021-11-11 15:07:52 <imode       > who wants to get forth up and running on a turing machine.
2021-11-11 16:13:10 <f-a         > I guess many ppl
2021-11-11 17:32:29 <lispmacs[work]> siraben: I finally found a TI-84+ calculator, and restored it to working order
2021-11-11 17:32:47 <lispmacs[work]> $6 at the local second-hand store
2021-11-11 17:33:06 <f-a         > pretty good price, I say
2021-11-11 17:33:08 <lispmacs[work]> I wanted to try out ti84-forth on it
2021-11-11 17:33:30 <lispmacs[work]> f-a: the only catch is it is the pink cover model, and pink is definitely not my color
2021-11-11 17:34:06 <lispmacs[work]> the pink case and pink cover are detachable though, if I could find spare parts from a black one
2021-11-11 17:34:50 <f-a         > maybe you could even paint it
2021-11-11 17:34:59 <Zarutian_HTC> how sought after are these pink parts?
2021-11-11 17:35:10 <f-a         > yeah you sould even sell them for a profit
2021-11-11 17:35:10 <lispmacs[work]> siraben: this TI-84+ appears to be a "silver edition" model. does that make a difference for ti84-forth?
2021-11-11 17:35:46 <Zarutian_HTC> i am thinking there might be an ti enthusiast willing to trade
2021-11-11 17:36:11 <lispmacs[work]> I think I'll just email the TI users list and see if anybody wants to trade
2021-11-11 17:36:56 <f-a         > do it!
2021-11-11 19:39:15 <MrMobius    > the numbering on the 84s is very confusing
2021-11-11 19:39:30 <MrMobius    > the wikipedia article is a good reference
2021-11-11 19:40:05 <MrMobius    > some of the 84s have a different cpu and os