2021-12-11 13:31:12 <veltas      > larsb: You can find the manual online for free, it's on the greenarrays website I think
2021-12-11 13:31:15 <veltas      > But just google it
2021-12-11 13:31:27 <veltas      > As for the software I don't know
2021-12-11 15:17:12 <remexre     > "traditionally," does the address of the end of a DO loop become part of the loop-sys (to use forth-standard.org's terminology)?
2021-12-11 15:19:09 <remexre     > previously I've just had do loops use a separate (software) stack, since nesting more than 3 in a single word is awful
2021-12-11 15:19:23 <remexre     > but it seems like I shouldn't need that
2021-12-11 15:49:07 <larsb       > Well well! http://www.bitsavers.org/pdf/hp/terminal/2392/firmware/
2021-12-11 15:49:16 <larsb       > Sorry, wrong channel.
2021-12-11 15:49:25 <larsb       > And yes, I saw the manual.
2021-12-11 16:00:29 <f-a         > Well well! https://twitter.com/TheASF/status/1400875147163279374
2021-12-11 16:00:32 <f-a         > Sorry, wrong channel.
2021-12-11 16:00:33 <f-a         > ( ͡° ͜ʖ ͡°)
2021-12-11 17:18:46 <veltas      > remexre: fig-forth's LEAVE would just set I to the limit, and not exit immediately
2021-12-11 17:19:10 <veltas      > I don't know about the old forth Inc approach
2021-12-11 17:19:50 <remexre     > huh, the forth-standard.org definition wouldn't allow that behavior; it says the loop is exited immediately
2021-12-11 17:20:01 <veltas      > The ANS behavior is leave will also jump out of the loop, which is where the third item comes from
2021-12-11 17:21:31 <veltas      > It's not necessary to have it there, you can keep a list of LEAVE's (and which do...loop level they're at) while compiling so every LEAVE can be matched up properly on encountering the corresponding LOOP
2021-12-11 17:21:54 <veltas      > But this is waaaay more complicated than any other core word's compilation
2021-12-11 17:22:00 <remexre     > yeah, but that requires an extra stack to make IF LEAVE THEN work, right?
2021-12-11 17:22:05 <veltas      > Yep
2021-12-11 17:22:10 <remexre     > yeah, I'll probably just do a 3-item loop-sys
2021-12-11 17:22:25 <veltas      > That's what I did in zenv
2021-12-11 17:22:51 <veltas      > In the one I'm writing now I think I'll keep the fig forth behavior
2021-12-11 17:23:02 <veltas      > It's good enough 99% of the time and way simpler
2021-12-11 17:23:42 <tabemann    > in my zeptoforth I opted for the jump-out-of-loop behavior
2021-12-11 17:24:47 <remexre     > yeah, I'm planning on doing an inlining+optimizing metacompiler at "some point," so I'll probably do the expensive thing until then
2021-12-11 17:24:53 <remexre     > because then it won't be any more expensive
2021-12-11 17:27:12 <veltas      > But yeah this is probably one of the jankiest things in the core words
2021-12-11 17:29:15 <veltas      > The state-smart version of S" is now core in forth200x
2021-12-11 17:29:56 <veltas      > ENVIRONMENT? is probably the worst core word and nobody will complain if it's missing
2021-12-11 17:32:05 <tabemann    > I don't even bother with trying to follow ANS
2021-12-11 17:32:42 <tabemann    > of course zeptoforth is an embedded Forth and it's not intended to run programs aimed at PC's
2021-12-11 17:34:23 <veltas      > Having a 2-item loop-sys means you can have I I' J be the first three return stack items
2021-12-11 17:46:25 <tabemann    > that doesn't really provide any benefit to me, because I've implemented DO-LOOPs in assembly, and I use SP-relative loads and stores for accessing the return stack in loops
2021-12-11 21:54:29 <concrete-houses> how good is forth for cgi and web pages and dynamic web content?
2021-12-11 21:54:38 <concrete-houses> I think I saw something called 1991....
2021-12-11 21:54:43 <concrete-houses> tries to rememeber..
2021-12-11 22:11:51 <remexre     > concrete-houses: https://www.1-9-9-1.com/
2021-12-11 22:12:51 <remexre     > I haven't used it tho; in general, I've avoided doing string processing tasks in Forth until I can get around to writing a regex->dfa compiler and lalr pgen in it
2021-12-11 23:06:06 <concrete-houses> interesting