2026-06-23 00:00:13 <lofty       > though my CTO's CPU is showing its age
2026-06-23 00:01:17 <tabemann    > your CTO created PicoRV32?
2026-06-23 00:01:43 <lofty       > and the B extension 0.9 proposal
2026-06-23 00:06:46 <tabemann    > forgive my ignorance, but is PicoRV32 a soft core?
2026-06-23 00:06:52 <lofty       > yes
2026-06-23 00:06:57 <veltas      > Is B a collection of other extensions?
2026-06-23 00:07:24 <tabemann    > (I get the impression that these days, if someone decides "hey I want to make a CPU", they implement RV32I)
2026-06-23 00:07:26 <lofty       > was written as a demo for icestorm to put on an iCE40HX1K
2026-06-23 00:08:08 <lofty       > Is B a collection of other extensions? <--- eeeh, originally was one extension, then got split up, and then only a few bits of the 0.9 proposal got standardised
2026-06-23 00:08:39 <tabemann    > it's kind of to CPU's what Forth is to computer languages (so many implementations to choose from!)
2026-06-23 00:09:49 <veltas      > Okay
2026-06-23 00:10:17 <lofty       > a full B proposal would make RV state of the art for bit manipulation
2026-06-23 00:10:24 <lofty       > *the full B proposal
2026-06-23 00:10:37 <lofty       > so naturally it languished in the committee stage and got picked apart
2026-06-23 00:11:05 <tabemann    > so why didn't they go with it? the fear of making something that could not be whittled down to a minimal toy/1c MCU implementation?
2026-06-23 00:12:21 <lofty       > ...kinda, yeah. I mean, Claire wrote an actual implementation of all the logic in the B v0.9 proposal to show how little logic it actually uses
2026-06-23 00:12:51 <lofty       > sometimes I call this the "you ain't gonna need it" vs "make the most of it" debate
2026-06-23 00:13:45 <tabemann    > I think one thing RV really suffers from is the tendency to not standardize useful things because someone might want to make a 1c MCU without them
2026-06-23 00:14:04 <tabemann    > like the interrupt model, or the lack of a decent PUSH/POP
2026-06-23 00:14:50 <lofty       > did you ever come across the term "macroinstruction fusion"? some of the RV people have talked a lot about it
2026-06-23 00:15:16 <tabemann    > I've heard the term, and it seems like a stupid way to work around an overly-limited ISA
2026-06-23 00:15:42 <tabemann    > can someone kickban cleobuline for a day?
2026-06-23 00:16:24 <xentrac     > don't be mean
2026-06-23 00:16:38 <lofty       > I think the worst bit is that RV makes a really bad instruction set for macroinstruction fusion
2026-06-23 00:16:52 <tabemann    > xentrac, who else has been sometimes joining and quitting multiple times within a minute?
2026-06-23 00:19:03 <tabemann    > when I was working on zeptoforth-V I was really getting hit hard by the fact that I couldn't use the PUSH/POP extension because A) I wanted to write portable code, B) I didn't want to infest my code with lots of 'minicode', and C) I didn't want to be limited by 4-word SP alignment
2026-06-23 00:19:26 <tabemann    > okay, I didn't write that right
2026-06-23 00:19:35 <tabemann    > 'minicode' isn't related to PUSH/POP
2026-06-23 00:19:41 <lofty       > because the obvious thing to fuse would be 32-bit immediates, to me. but...LUI and AUIPC take a register as destination, so you need an extra register file write port to handle fusion patterns more generally
2026-06-23 00:19:47 <xentrac     > millicode
2026-06-23 00:19:52 <cleobuli_   > hello
2026-06-23 00:19:54 <tabemann    > yeah
2026-06-23 00:20:11 <tabemann    > cleobuli_, you've been repeatedly joining and parting multiple times today, often within the same minute
2026-06-23 00:20:38 <tabemann    > are you using IRC on a cell phone or something?
2026-06-23 00:21:21 <tabemann    > xentrac, that's it
2026-06-23 00:22:17 <xentrac     > lofty: which one was B?
2026-06-23 00:22:30 <lofty       > Bit manipulation
2026-06-23 00:22:36 <tabemann    > isn't the current standardized B extension a small set of bit operations?
2026-06-23 00:23:13 <lofty       > https://github.com/riscv/riscv-bitmanip/releases/tag/v0.93
2026-06-23 00:25:59 <tabemann    > lofty: it's sad that they didn't go with that
2026-06-23 00:26:22 <lofty       > when I called it "state of the art" I was not kidding :p
2026-06-23 00:27:18 <tabemann    > but of course committees have to screw things up, as they always do
2026-06-23 00:28:33 <xentrac     > tabemann: SP alignment is a pretty bad fit for Forths
2026-06-23 00:30:06 <tabemann    > xentrac, precisely
2026-06-23 00:30:22 <tabemann    > especially if you're on a microcontroller where you can't afford to waste SRAM
2026-06-23 00:31:14 <xentrac     > I think the tradeoffs around interrupt models are somewhat unavoidable
2026-06-23 00:31:32 <tabemann    > I'm spoiled by the ARM Cortex-M interrupt model
2026-06-23 00:31:54 <tabemann    > and I designed zeptoforth around it
2026-06-23 00:32:19 <tabemann    > which is why I'm stuck on zeptoforth-V
2026-06-23 00:32:26 <xentrac     > everything needs either extra chip real estate or extra time, and the cost ratio of those things varies by orders of magnitude
2026-06-23 00:33:06 <xentrac     > usually interrupt handling is a pretty small part of a system like zeptoforth; how did it end up being pervasive in the way you are suggesting?
2026-06-23 00:33:35 <tabemann    > the problem with RV is that they made choices that make sense for 1c MCU's that make no sense for anything else, while simultaneously making choices that make sense for application processors that make little sense for anything else
2026-06-23 00:34:16 <tabemann    > xentrac, the problem is that I'm trying to make a portable interrupt model like that for the original zeptoforth
2026-06-23 00:34:44 <tabemann    > but I need to access certain RP2350 interrupts in the kernel, which have the effect of breaking that portability
2026-06-23 00:35:15 <tabemann    > in the original zeptoforth on the RP2350 all I did was set a vector handler for a particular IRQ and all was good with the world
2026-06-23 00:35:45 <tabemann    > whereas with zeptoforth-V I have to effectively break portability in the kernel itself in a much more fundamental way
2026-06-23 00:36:22 <tabemann    > (the particular issue is that I need to capture inter-CPU communication interrupts in the kernel so I can trap the PC in SRAM when doing things like programming/erasing flash)
2026-06-23 00:38:46 <tabemann    > after all, my intention with zeptoforth-V was to write a general RV32I (and possibly RV64I, I made some choices deliberately to make it easy to port to RV64) Forth, but it seems like I might end up having to bake a lot of unportable functionality that I'm trying to avoid into the kernel itself (beyond obviously unportable stuff like hardware initialization and flash drivers)
2026-06-23 00:39:17 <xentrac     > btw, there aren't currently any 1¢ MCUs
2026-06-23 00:39:47 <tabemann    > "1c" was hyperbole
2026-06-23 00:40:07 <xentrac     > it's not very hyperbolic.  the cheapest ones are 2¢, in quantity 600: https://www.lcsc.com/product-detail/C49173938.html
2026-06-23 00:40:31 <xentrac     > down to 1.77¢ in quantity 9000
2026-06-23 00:42:48 <tabemann    > that's $159.30 for 9000 MCU's, which sounds ridiculous
2026-06-23 00:42:58 <tabemann    > usually when you buy that may chips you expect to pay more than that
2026-06-23 00:43:03 <tabemann    > *many
2026-06-23 00:43:29 <xentrac     > isn't that insane?
2026-06-23 00:43:37 <xentrac     > they're SOT-23-6s though
2026-06-23 00:44:18 <xentrac     > so probably about 10mg each, so 9000 computers should weigh about 90 grams
2026-06-23 00:44:26 <tabemann    > I presume these have like one or two GPIO's tops
2026-06-23 00:44:30 <xentrac     > 4
2026-06-23 00:44:47 <xentrac     > I think the cheapest RISC-V is probably the CH32V003 in a SOP-8: https://www.lcsc.com/product-detail/C5346354.html
2026-06-23 00:45:23 <xentrac     > which is like 11¢ or 12¢
2026-06-23 00:45:48 <tabemann    > $423.60 for 4000 of them, which is still damn cheap
2026-06-23 00:45:53 <xentrac     > yup
2026-06-23 00:46:22 <xentrac     > though to my eye the more staggering number is US$62.20 for 500 of them
2026-06-23 00:46:45 <tabemann    > I've seen *single* MCU dev boards that cost more than twice that
2026-06-23 00:47:04 <xentrac     > I'd like 500 32-bit RISC workstations please.  Oh, payment?  I think I have some coins in my pocket, let me check.
2026-06-23 00:48:29 <tabemann    > 2K SRAM? still more than the amount of RAM in an Atari VCS
2026-06-23 00:48:42 <xentrac     > yeah, 16× as muc
2026-06-23 00:48:42 <xentrac     > h
2026-06-23 00:49:40 <tabemann    > the only reason to not get these rather than, say, the venerable 8051 for new designs is who knows when the chip will be shelved as CH moves onto other things
2026-06-23 00:50:01 <xentrac     > yeah, although if you're concerned about that, maybe go with an STM32
2026-06-23 00:50:01 <tabemann    > (whereas you know the 8051 isn't going anywhere)
2026-06-23 00:50:32 <xentrac     > that reminds me, I dissected a DVD player and found out that its main CPU was an 8051
2026-06-23 00:50:47 <tabemann    > was it one of those super-8051's though?
2026-06-23 00:50:51 <xentrac     > a humongous system-on-chip which included an on-chip MPEG decoder, yeah
2026-06-23 00:51:06 <tabemann    > I thought so
2026-06-23 00:51:28 <xentrac     > so that's the other kind of case where pairing a hardwired CPU with an FPGA could maybe save you power
2026-06-23 00:51:51 <xentrac     > an MPEG decoder in an FPGA might use a lot more power than a hardwired MPEG decoder
2026-06-23 00:52:05 <xentrac     > but it's likely to use less power than an MPEG decoder on the CPU
2026-06-23 00:52:11 <tabemann    > yeah
2026-06-23 00:52:29 <xentrac     > and now that you can easily buy hardwired MPEG decoders, the argument switches to something like AV1 decoders or something
2026-06-23 00:52:31 <tabemann    > but if you're making DVD players you *could* probably afford to commission an ASIC
2026-06-23 00:52:44 <xentrac     > historically true, probably not now
2026-06-23 00:53:07 <tabemann    > yeah, things have marched onto streaming media
2026-06-23 00:53:30 <tabemann    > thing is, why would you try to save power in a DVD player?
2026-06-23 00:53:43 <tabemann    > you're probably hooked up to a power brick hooked up to mains
2026-06-23 00:54:13 <tabemann    > the obvious motivation would be to save money, because these days DVD players aren't exactly high-margin items
2026-06-23 00:54:30 <xentrac     > nowadays you're probably decoding the video on your cellphone on the subway
2026-06-23 00:54:52 <tabemann    > your cell phone probably has a hardware MPEG decoder built in
2026-06-23 00:55:24 <xentrac     > yes, and an H.264 decoder, but not an AV1 decoder or a VP8 decoder
2026-06-23 00:55:41 <tabemann    > yeah
2026-06-23 00:58:24 <xentrac     > I've also been amazed at the prices of some other electronic components on LCSC.  3¢ piezo buzzer: https://www.lcsc.com/product-detail/C22387762.html  1.6¢ headphone jack: https://www.lcsc.com/product-detail/C52750867.html  1.4¢ 6.5GHz transistor: https://www.lcsc.com/product-detail/C42370039.html  3¢ 74HC595: https://jlcpcb.com/partdetail/HGSEMI-HG74HC595MTR/C19193503  2¢ LM358 dual op-amp upgraded
2026-06-23 00:58:30 <xentrac     > to 4MHz: https://www.lcsc.com/product-detail/C53373913.html
2026-06-23 00:58:48 <tabemann    > hmm... apparently there is such a thing as mobile phone hardware VP8 support
2026-06-23 00:59:11 <xentrac     > oh, cool
2026-06-23 00:59:27 <tabemann    > yeah, apparently there are designs including it dating back as far as 2013
2026-06-23 01:01:57 <xentrac     > wow, I'm pretty out of the loop!
2026-06-23 01:02:28 <tabemann    > older cell phones implemented it in software though
2026-06-23 01:02:49 <tabemann    > (VP8 came out in 2008)
2026-06-23 01:04:21 <tabemann    > okay, I'm going to have dinner now, so I'll bbl
2026-06-23 01:04:34 <xentrac     > talk to you later!
2026-06-23 06:10:37 <Stalevar    > Chuck Moor said he updated Windows to 10 or 11 and ColorForth stopped working. I wonder why Windows on the first place? If he's a hacker it doesn't make sense to use windows
2026-06-23 06:12:56 <Stalevar    > Though you can spell it 04color03Forth in IRC
2026-06-23 06:14:59 <Stalevar    > https://www.reddit.com/r/Forth/comments/1ozsajc/windows_update_break_colorforth_chuck_thinks/ (contains youtube video MvkGBWXb2oQ in frame)
2026-06-23 08:40:50 <xentrac     > Stalevar: he was having a hard time writing drivers for new hardware
2026-06-23 08:41:33 <Stalevar    > Of course. That's where his philosophy bit him, I guess?
2026-06-23 08:42:09 <Stalevar    > In 60s, 70s, 80s and even 90s you can talk with hardware directly and get meaningful result
2026-06-23 08:42:39 <Stalevar    > But in 2010 onward where you can't even use RAM without proprietary training code? Not so much
2026-06-23 08:43:04 <xentrac     > well, there's that, and then there's the time iTV Corp. failed because Chuck's numerical simulation of the fab's processes was inaccurate in a way that produced a dud tapeout
2026-06-23 08:43:14 <xentrac     > so they ran out of money
2026-06-23 08:43:53 <xentrac     > more generally speaking they were pushing too hard to get to the optimum when they could have settled for "good enough"
2026-06-23 08:44:25 <Stalevar    > Last 10% of anything usually require 90% effort
2026-06-23 08:45:00 <xentrac     > also I suspect that omitting multipliers from the GA144 (and the F21/i21, and the MuP21) was a bad tradeoff; that ruled them out for the workloads that could have really benefited from the GA144's "unlimited supply of MIPS"
2026-06-23 08:45:47 <Stalevar    > I'm a little out of context. Moore tried to make gazillion core CPUs optimized to run Forth?
2026-06-23 08:45:56 <xentrac     > he succeeded
2026-06-23 08:46:15 <Stalevar    > Where were they used?
2026-06-23 08:46:35 <xentrac     > nowhere
2026-06-23 08:46:36 <xentrac     > http://www.greenarraychips.com/home/documents/greg/DB001-110412-F18A.pdf
2026-06-23 08:48:39 <Stalevar    > Then it doesn't sound like "succeeded"?
2026-06-23 08:48:47 <xentrac     > he made the chips
2026-06-23 08:49:04 <xentrac     > the GA144 has 144 cores which run about a billion instructions per second each
2026-06-23 08:49:05 <Stalevar    > But he can't even procure some to power his own workstation?
2026-06-23 08:49:24 <Stalevar    > and had to use windows 11?
2026-06-23 08:49:45 <xentrac     > well, he can get the chips, but that's just a CPU.  he doesn't have the peripherals
2026-06-23 08:50:16 <Stalevar    > What's the point of CPU if there is no motherboard to plug it into?
2026-06-23 08:50:30 <xentrac     > I think they sell evaluation boards actually
2026-06-23 08:50:43 <xentrac     > each GA144 CPU has 64 words of RAM, 18 bits each
2026-06-23 08:52:37 <xentrac     > and another 64 words of ROM.  each word holds up to 4 5-bit instructions, but the last instruction can only be one of 8 possibilities
2026-06-23 08:53:34 <Stalevar    > Sounds more like a specialized MCU or something rather than general purpose CPU
2026-06-23 08:54:38 <xentrac     > you could in theory use it that way, but a normal MCU would probably be a lot easier
2026-06-23 08:54:59 <xentrac     > some of the CPUs around the edge of the array control a single GPIO each
2026-06-23 08:55:13 <Stalevar    > Still, if he's rich and smart enough to make his own chips, why not to make an entire computer to run 04color03Forth?
2026-06-23 08:55:31 <Stalevar    > Instead of struggling with Windows 10/11
2026-06-23 08:56:20 <Stalevar    > xentrac, so it's more like smart FGPA?
2026-06-23 08:58:01 <xentrac     > I'd like to know that too.  you'd think the GA144 wouldn't struggle with producing a video signal
2026-06-23 08:58:18 <xentrac     > or decoding the PS/2 keyboard protocol
2026-06-23 08:58:33 <xentrac     > it's in this uncomfortable space between FPGAs and normal CPUs
2026-06-23 08:59:02 <xentrac     > where you can't practically program it in either Verilog or C
2026-06-23 08:59:33 <xentrac     > but instead must manually lay out a floorplan and then program each node in MachineForth
2026-06-23 09:00:25 <Stalevar    > There is a game (sadly proprietary) OpusMagnum, where you have to arrange arms and program them to build "molecules". I guess programming that machine would be a similar struggle
2026-06-23 09:00:34 <iv4nshm4k0v > Sounds a bit like IA-64: an interesting concept, but no devtools ever appeared to make good use of it.
2026-06-23 09:00:36 <Stalevar    > Maybe there are free games similar to it
2026-06-23 09:00:49 <xentrac     > I don't think he's rich.  the settlement with Leckrone was apparently enough to pay for getting the GreenArrays chips out the door, but he lives in a log cabin in the woods
2026-06-23 09:04:43 <xentrac     > last I checked, you could tape out a prototype chip with an "MPW"/"shuttle" service like MOSIS or CMP for about US$4000
2026-06-23 09:05:56 <xentrac     > and now we have Tiny Tapeout, which reduces that to US$300 but is indeed pretty tiny (because you're sharing a chip with a lot of other designs)
2026-06-23 09:08:02 <xentrac     > I don't know why Chuck didn't have his own GA144-based workstation.  bitbanging a video signal is not rocket science, and his MuP21 included NTSC generation hardware, so he's familiar with what they consist of
2026-06-23 09:23:48 <Stalevar    > What would you pick out of pForth, SP-Forth, Lina, GForth or Jones Forth (though probably not useable as actual engine even if it works)
2026-06-23 09:24:02 <Stalevar    > or what else runs on Linux natively?
2026-06-23 09:36:00 <veltas      > I tend to use gforth, despite my issues with it
2026-06-23 09:39:22 <veltas      > xentrac: Maybe Chuck was burned out
2026-06-23 09:43:38 <Stalevar    > What issues?
2026-06-23 10:01:30 <xentrac     > veltas: plausible
2026-06-23 10:02:13 <xentrac     > I've tried Gforth, pForth, and PFE.  of those three I like Gforth best
2026-06-23 10:02:33 <xentrac     > not familiar with Lina
2026-06-23 10:04:08 <xentrac     > for https://canonical.org/~kragen/sw/dev3/wmaze.fs I had to stub out time&date, form, and at-xy to get it to run in PForth.  and I don't think that's because they're present under a different name; I think they're just missing functionality
2026-06-23 10:04:25 <xentrac     > similar with PFE IIRC, though I don't think I got wmaze.fs running in it
2026-06-23 10:07:46 <Stalevar    > There is also zForth but it's minimal
2026-06-23 10:18:48 <veltas      > Issues with gforth: they've not released in well over a decade, even though there's a lot of activity, and it's trying to be too much like a normal programming language with its warnings and the bloat
2026-06-23 10:19:14 <veltas      > That said it's got a lot of useful features like C integration and conforms to ANS standard well enough
2026-06-23 10:20:03 <veltas      > My biggest complaint is they haven't bothered doing a release in so long, most people using gforth are using a version that's far removed from what the devs are using
2026-06-23 10:20:49 <xentrac     > Debian seems to have dropped GForth
2026-06-23 10:21:37 <veltas      > The draft standard would probably do better if gforth was being released, they're one of the only big forths that is open source and follows the draft spec
2026-06-23 10:27:11 <Stalevar    > I like stack underflow detection
2026-06-23 10:27:41 <Stalevar    > Jones Forth spit some numbers if you keep taking numbers from empty stack and who knows what happens if you overwrite them
2026-06-23 10:28:11 <veltas      > xentrac: Found a little context https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1039051 https://tracker.debian.org/news/1730944/accepted-gforth-073dfsg-10-source-into-unstable/
2026-06-23 10:29:30 <veltas      > Stack under/overflow is trivial to add, only very minimalist forths lack it
2026-06-23 10:29:41 <veltas      > Admittedly I don't think zenv has it, maybe I should add it
2026-06-23 10:30:16 <veltas      > xentrac: So sounds like Bernd (gforth leader) will get 0.7.3 back in for next release of Debian
2026-06-23 10:30:23 <Stalevar    > Or you can just put stack on an edge of permitted memory then undeflow will be a hardware exception
2026-06-23 10:30:42 <veltas      > I think that's how gforth does it(?)
2026-06-23 10:30:43 <xentrac     > veltas: that's good!
2026-06-23 10:30:45 <lofty       > 0.7.3 is still, like, 2008 though, right?
2026-06-23 10:31:03 <veltas      > Gforth catches signals like segfaults and returns to interpreter
2026-06-23 10:31:13 <xentrac     > yeah, that's what you'd want, isn't it?
2026-06-23 10:31:19 <veltas      > It's actually relatively easy to do this although KipIngram knows some of the weeds
2026-06-23 10:31:22 <xentrac     > that way you don't lose your work
2026-06-23 10:31:27 <veltas      > Of course
2026-06-23 10:31:27 <Stalevar    > Yes, but in Slackware I just did ./configure --prefix=/opt/gforth and make and make install and it worked
2026-06-23 10:31:39 <veltas      > For an 'environment' rather than just a 'program' it makes sense yes
2026-06-23 10:31:49 <xentrac     > oh, I thought this was one of the things on the list of "my issues with [GForth]"
2026-06-23 10:31:54 <veltas      > No :P
2026-06-23 10:31:59 <veltas      > We're just chatting now
2026-06-23 10:32:05 <xentrac     > sorry!
2026-06-23 10:32:15 <veltas      > My issue is the stable release is from 'like, 2008'
2026-06-23 10:32:40 <Stalevar    > If it works why change?
2026-06-23 10:32:40 <veltas      > We need to have a big bash for 0.7.3's 20th
2026-06-23 10:32:55 <veltas      > Everyone's invited
2026-06-23 10:33:36 <xentrac     > yes!
2026-06-23 10:33:40 <veltas      > Stalevar: They have changed it loads and are on unreleased version 0.7.9, which is more like 0.9.3 or whatever because it's totally different model internally
2026-06-23 10:33:48 <Stalevar    > Anyway, why not to try SP-Forth or https://home.hccnet.nl/a.w.m.van.der.horst/lina.html
2026-06-23 10:34:07 <veltas      > But they haven't done a 'release' since 2008 so people are stuck on an old unloved version
2026-06-23 10:34:34 <veltas      > 3 years ago they said they were near to 1.0, which I think is what they're holding off for
2026-06-23 10:35:18 <veltas      > Their criteria for 1.0 doesn't sound worth it, I think they should release now.  I think they want better documentation first
2026-06-23 10:36:22 <xentrac     > this reminds me of people with a hoarding problem
2026-06-23 10:37:03 <xentrac     > by the way, right now my kitchen counter is full of parts from a microwave I just gutted
2026-06-23 10:37:15 <veltas      > In a way, writing a Forth is a bit like a hoarding problem
2026-06-23 10:38:06 <xentrac     > oh?
2026-06-23 10:39:21 <veltas      > Well you're collecting and spending time surrounding yourself with stuff that's junk to most people, and is preventing you from a normal healthy lifestyle
2026-06-23 10:39:39 <veltas      > A lot of computing hobbies are worse than smoking
2026-06-23 10:39:50 <xentrac     > stuff such as UM/MOD?
2026-06-23 10:39:56 <veltas      > Yeah exactly
2026-06-23 10:40:19 <veltas      > Damn you went straight for my heart there :P
2026-06-23 10:40:52 <xentrac     > I think um/mod is brilliant but it is a rather esoteric problem
2026-06-23 10:41:08 <xentrac     > as I was breathing the smoke from microwave circuit boards last night I was wondering how healthy it was
2026-06-23 10:41:11 <veltas      > Isn't UM/MOD just DIV
2026-06-23 10:41:43 <xentrac     > it's divmod
2026-06-23 10:41:53 <veltas      > I mean DIV the x86 instruction
2026-06-23 10:42:08 <xentrac     > oh, I thought you meant the Pascal operator
2026-06-23 10:42:26 <veltas      > It can be implemented as the one-parameter DIV instruction
2026-06-23 10:42:27 <xentrac     > yes, it's exactly DIV the x86 instruction
2026-06-23 10:42:56 <xentrac     > if I understand them both correctly
2026-06-23 10:42:59 <veltas      > UM/MOD is actually quite fundamental and I think most platforms have it
2026-06-23 10:43:09 <veltas      > What's weird is what e.g. C doesn't have it
2026-06-23 10:43:24 <veltas      > Well it's not weird because C is a systems language and not a portable assembly
2026-06-23 10:43:24 <xentrac     > C doesn't really have mixed-mode arithmetic
2026-06-23 10:43:38 <Stalevar    > xentrac, the difference is you can set up ventilation or put on a respirator. but if you are smoking you can't
2026-06-23 10:44:09 <Stalevar    > But yeah, breathing toxic fumes regularly is probably as bad as smoking
2026-06-23 10:44:21 <veltas      > Could easily be worse than smoking
2026-06-23 10:44:36 <Stalevar    > Or better
2026-06-23 10:44:46 <veltas      > Smoking's not good for you but I think that people often end up with a hysterical view of it
2026-06-23 10:44:48 <xentrac     > Stalevar: I think the main health problem with computing hobbies is just that you spend a lot of time sitting down
2026-06-23 10:45:24 <Stalevar    > By the way starting forth epub from web archive is so broken that it won't open in FBreader
2026-06-23 10:45:26 <xentrac     > yes, probably the guy who smokes a cigar every other weekend is not going to have much of a health effect from it
2026-06-23 10:45:31 <Stalevar    > Though Book Reader kinda worked
2026-06-23 10:45:56 <Stalevar    > xentrac, it's an exception. Most of smokers smoke every day and more than once
2026-06-23 10:46:08 <veltas      > The rate of lung cancer for people who smoke like 20 cigarettes a day is like 14% in their whole life, that is way lower than I would have guessed without knowing that statistic
2026-06-23 10:46:25 <veltas      > I can't remember exact numbers but anyone can go look this up, I'm sure someone will google this outrageous statement
2026-06-23 10:46:33 <Stalevar    > If one smokes every other weekend it might not be too bad, but it's a very rare exception, most will either smoke several times a day or not at all
2026-06-23 10:46:50 <xentrac     > veltas: I thought it was a little lower than that, but maybe I'm thinking of the lung cancer rate for all smokers
2026-06-23 10:46:53 <veltas      > 20 a day is a ridiculous amount although sadly many people smoke that much
2026-06-23 10:47:01 <xentrac     > yeah, it's common
2026-06-23 10:47:51 <Stalevar    > I don't see a point in smoking. Nicotine doesn't even give that much of a kick as I heard
2026-06-23 10:48:16 <veltas      > I think it's more that they're addicted so it calms their withdrawal
2026-06-23 10:48:26 <veltas      > Rather than it having inherent calming properties
2026-06-23 10:48:34 <xentrac     > Stalevar: https://gwern.net/nicotine
2026-06-23 10:49:28 <Stalevar    > also regarding mecrisp, I'm a little confused with all stellaris, quarta and so on. Which one I need?
2026-06-23 10:49:38 <Stalevar    > for native linux if it works at all
2026-06-23 10:50:02 <veltas      > I wonder if steroid abuse is worse
2026-06-23 10:50:45 <veltas      > Apparently that can cause cancer too and it's getting common with body builders
2026-06-23 10:50:48 <xentrac     > not sure about Mecrisp either
2026-06-23 10:50:57 <veltas      > Steroids / testosterone
2026-06-23 10:51:04 <xentrac     > steroid abuse is potentially a lot worse than cancer
2026-06-23 10:51:17 <veltas      > Stalevar: I don't think mecrisp runs on native linux
2026-06-23 10:51:18 <xentrac     > not many things are, but steroids can be
2026-06-23 10:51:34 <veltas      > Mecrisp is meant for MCUs, the different named ones run on different platforms / archs
2026-06-23 10:52:03 <veltas      > Like Mecrisp Quintus is for RISC-V
2026-06-23 10:52:43 <veltas      > I think if you take a bunch of steroids every day for 10 years that has a super high mortality rate
2026-06-23 10:52:50 <veltas      > Way higher than cigarettes
2026-06-23 10:52:51 <xentrac     > it doesn't take 10 years
2026-06-23 10:53:03 <veltas      > Well it's all statistics right
2026-06-23 10:53:06 <xentrac     > the reason most anti-inflammatory drugs are called "non-steroidal anti-inflammatory drugs" is the iatrogenic catastrophe that ensued upon the initial discovery of steroids: https://www.researchgate.net/publication/264248818_Origins_and_impact_of_the_term_%27NSAID%27
2026-06-23 10:54:41 <Stalevar    > mecrisp-2.0.7.tar.gz is mecrisp-stellaris?
2026-06-23 10:55:09 <veltas      > Stalevar: there's #mecrisp on hackint.org IRC
2026-06-23 10:55:21 <veltas      > And also libera.chat but I think that's less active
2026-06-23 10:55:32 <xentrac     > steroid therapy started in 01949, won its discoverers the Nobel Prize in 01950, and patients started dying horribly en masse around 01953
2026-06-23 10:55:55 <xentrac     > of course the dose makes the poison
2026-06-23 10:55:55 <veltas      > Wow
2026-06-23 10:56:06 <veltas      > A lot of early Nobel prizes associated with lots of people dying
2026-06-23 10:56:26 <xentrac     > well, those steroid therapies are still in use today
2026-06-23 10:56:34 <xentrac     > they just try to strictly limit their duration
2026-06-23 10:57:26 <Stalevar    > Did you know that they would inject thorium oxide suspense in vein?
2026-06-23 10:57:44 <veltas      > I think 'mecrisp' is for MSP430
2026-06-23 10:57:48 <xentrac     > no, I only knew about it in intestines
2026-06-23 10:57:51 <Stalevar    > Yeah, that thorium people panic over a tiny amount in rods
2026-06-23 10:57:51 <veltas      > Mecrisp Stellaris is for ARM
2026-06-23 10:58:01 <veltas      > I can't tell you more than google though
2026-06-23 10:58:09 <veltas      > I've only used Mecrisp once
2026-06-23 10:58:18 <Stalevar    > Alright, I'll just use stellaris then
2026-06-23 10:58:32 <xentrac     > I wonder if it should have been Mecrisp Brachius
2026-06-23 10:58:34 <Stalevar    > It has a nice readme listing forth words one per line, which just what I needed
2026-06-23 10:58:46 <veltas      > xentrac: Is that a joke?  It might be over my head
2026-06-23 10:59:01 <xentrac     > "brachium" is Latin for "arm"
2026-06-23 10:59:03 <veltas      > Mecrisp is a great project
2026-06-23 10:59:15 <veltas      > Very good xentrac
2026-06-23 11:00:42 <veltas      > The author of mecrisp is one of those people who just has loads of useful output and mostly keeps to themselves
2026-06-23 11:00:43 <xentrac     > total cigarette mortality rate among 20-a-day smokers is probably significantly higher than the 14% lung cancer rate, even though lung cancer isn't always fatal
2026-06-23 11:02:11 <xentrac     > https://en.wikipedia.org/wiki/Health_effects_of_smoking_tobacco mentions pneumonia, heart attacks, strokes, emphysema, chronic bronchitis, peripheral arterial disease, and high blood pressure
2026-06-23 11:02:53 <xentrac     > saying, "About one-half of men who smoke long-term will die of illness due to smoking."
2026-06-23 11:03:13 <veltas      > You're absolutely right about that, it is a little dishonest to say what I said
2026-06-23 11:03:21 <veltas      > In that it's a bit misleading
2026-06-23 11:03:59 <veltas      > But I will also say that I question how many of these people with "illness due to smoking" are really dying because of smoking or have had lifespans shortened significantly because of smoking
2026-06-23 11:04:12 <xentrac     > I didn't think it was dishonest; I just thought it was only part of the picture
2026-06-23 11:04:17 <veltas      > Nevertheless I don't smoke and don't plan on taking it up
2026-06-23 11:04:29 <veltas      > I think if you present things out of context it can be a little dishonest
2026-06-23 11:05:02 <Stalevar    > Does 0<> have any use? IF WHILE already accept numbers as true if they are not zero and false if zero?
2026-06-23 11:05:08 <veltas      > It adds nuance still to the conversation that lung cancel isn't a certainty
2026-06-23 11:05:35 <veltas      > Stalevar: It's a bool cast basically, so if you want to store a clean 'flag' you can use it
2026-06-23 11:05:40 <veltas      > Otherwise it's useless yes
2026-06-23 11:05:42 <xentrac     > lung cancer definitely isn't a certainty, and neither is death due to smoking
2026-06-23 11:05:58 <veltas      > Stalevar: I guess also it's useful as documentation that you're checking for a value not being zero
2026-06-23 11:06:09 <veltas      > But in practice most forthers don't bother with that kind of code
2026-06-23 11:06:45 <veltas      > 'lung cancel' I certainly don't want to get cancelled by my lungs
2026-06-23 11:06:54 <xentrac     > #metoo
2026-06-23 11:07:06 <xentrac     > I was amused to learn today that the proof assistant Lean has a new-to-me definition of inequality
2026-06-23 11:07:38 <xentrac     > in Lean `A ≠ B` is syntactic sugar for `A = B → False`
2026-06-23 11:07:54 <xentrac     > that is to say, that A = B would imply False
2026-06-23 11:08:10 <xentrac     > that if it were true that A equaled B, then False would be true
2026-06-23 11:09:06 <lofty       > and in python 2, it can be /j
2026-06-23 11:09:50 <xentrac     > I think dividing by 1j in Python is equivalent to multiplying by -1j
2026-06-23 11:10:54 <xentrac     > dividing by j isn't especially well defined because j is just a variable
2026-06-23 11:11:31 <xentrac     > (but not in Forth!)
2026-06-23 11:14:34 <veltas      > lofty: In Forth it can be
2026-06-23 11:15:09 <lofty       > but such would not be standard; in python the standard is "whatever cpython does" /lh
2026-06-23 11:15:27 <veltas      > TRUE CONSTANT FALSE  is entirely standard
2026-06-23 11:16:14 <lofty       > > A true flag returned by a standard word shall be a single-cell value with all bits set. A false flag returned by a standard word shall be a single-cell value with all bits clear.
2026-06-23 11:17:30 <MrMobius    > Mecrisp Across is msp430. Mecrisp Stellaris is ARM.
2026-06-23 11:18:28 <veltas      > lofty: Yes but I mean you can write a new definition for 'FALSE', not change what true/false stuff is accepted by existing words
2026-06-23 11:42:33 <Stalevar    > You recommended me to read starting forth, I read it up to line where they say to try 15 spaces.   Gforth has this, Create spaces ( u -- ) bl 80 times
2026-06-23 11:42:33 <Stalevar    > DOES> ( u -- ) swap 0 max 0 ?DO  I' I - &80 min 2dup type  +LOOP  drop ;
2026-06-23 11:43:56 <Stalevar    > So `create spaces` adds a dictionary entry, bl 80 times skips 80 bytes filling them with 0x32 and then finally DOES> define what happens when you call word SPACES ?
2026-06-23 11:44:38 <Stalevar    > But unlike regular action after colon, DOES> (or create?) add the address of created word to stack
2026-06-23 11:46:49 <Stalevar    > Then the loop spits 80 spaces at a time until there are less than 80, then it prints the line with remaining spaces using length argument to TYPE
2026-06-23 12:24:06 <Stalevar    > '\' $7F17A21E9FD8 c! \ok
2026-06-23 12:57:15 <Stalevar    > '\' ' prompt 160 + c! \ok
2026-06-23 12:58:54 <Stalevar    > forthBot: OK
2026-06-23 12:58:55 <forthBot    > Error: Unknown word: OK
2026-06-23 12:59:24 <Stalevar    > forthBot: ' prompt 198 dump
2026-06-23 12:59:24 <forthBot    > Error: Unknown word: '
2026-06-23 13:00:49 <Stalevar    > forthBot: LOAD ini.fth
2026-06-23 13:00:49 <forthBot    > File ini.fth with MOON  loaded
2026-06-23 13:00:56 <Stalevar    > forthBot: ' prompt 198 dump
2026-06-23 13:00:56 <forthBot    > Error: Unknown word: '
2026-06-23 13:08:43 <Stalevar    > forthBot: ['] prompt 198 dump
2026-06-23 13:08:43 <forthBot    > Error: Unknown word: [']
2026-06-23 13:09:41 <Stalevar    > forthBot: prompt
2026-06-23 13:09:42 <forthBot    > Error: Unknown word: prompt
2026-06-23 13:11:32 <lofty       > "when you've learned one forth, you've learned one forth" continues to be a truism.
2026-06-23 13:12:01 <Stalevar    > forthBot: : BREAK-STRING ( 0 ... n2 n1 addr -- ) BEGIN OVER WHILE TUCK ! 4 + REPEAT DROP ;
2026-06-23 13:12:43 <Stalevar    > forthBot: : MY-STRING S" This should be thirty--two bytes "S ;
2026-06-23 13:13:03 <Stalevar    > forthBot: MY-STRING TYPE CR
2026-06-23 13:13:03 <forthBot    > Error: Unknown word: TYPE
2026-06-23 13:13:25 <Stalevar    > forthBot: MY-STRING PRINT CR
2026-06-23 13:13:25 <forthBot    > This should be thirty--two bytes
2026-06-23 13:14:07 <Stalevar    > forthBot: : BREAK-STRING ( 0 ... n2 n1 addr -- ) BEGIN OVER WHILE TUCK ! 4 + REPEAT DROP DROP ;
2026-06-23 13:14:44 <Stalevar    > 0 31090 1869440365 543716457 1998615411 1701650548 661546092 1970235507 544567129 MY-STRING DROP BREAK-STRING MY-STRING PRINT CR
2026-06-23 13:14:58 <Stalevar    > forthBot: 0 31090 1869440365 543716457 1998615411 1701650548 661546092 1970235507 544567129 MY-STRING DROP BREAK-STRING MY-STRING PRINT CR
2026-06-23 13:14:58 <forthBot    > Error: STORE: invalid memory index 20756f59
2026-06-23 13:18:58 <Stalevar    > Wait, why https://home.hccnet.nl/a.w.m.van.der.horst/ci86.lina64-5.5.0.tar.gz is a zip archive containing file saying it's obsolete...
2026-06-23 13:19:02 <Stalevar    > https://home.hccnet.nl/a.w.m.van.der.horst/lina.html
2026-06-23 13:19:08 <Stalevar    > It's latest version isn't it?
2026-06-23 13:23:28 <vms14       > guys there is a cat lang game jam :D
2026-06-23 13:23:51 <vms14       > https://itch.io/jam/catjam-2026
2026-06-23 13:24:21 <vms14       > you could participate with forth
2026-06-23 13:24:35 <vms14       > i want to participate with my toy lang which i'm rewriting
2026-06-23 13:25:24 <vms14       > https://gitlab.com/ohmycat/oh/-/blob/main/oh.js
2026-06-23 13:26:21 <Stalevar    > vms14, explain what it is for those who cannot open the links
2026-06-23 13:27:21 <Stalevar    > God is King of kings and Lord of lords // Hm, which god?
2026-06-23 13:42:00 <vms14       > so you can open the links xd
2026-06-23 13:42:08 <vms14       > it's my toy lang
2026-06-23 13:42:40 <vms14       > it's a layer on top of js, everything in this lang is a js function
2026-06-23 13:42:49 <vms14       > and has two kind of immediate words, immediate 0 and immediate 1
2026-06-23 13:43:09 <vms14       > immediate 0 just execute, immediate 1 have to return a function on the stack, which will be appended to the runtime
2026-06-23 13:44:04 <vms14       > for example the line comment '---' is an immediate 0 word that discards the source code until newline or eof
2026-06-23 13:44:42 <vms14       > if is an immediate 1 word that compiles it's own code array and returns a closure that will execute the test block, and depending on what's on the stack after that, the true or false branch
2026-06-23 13:48:44 <Stalevar    > vms14, I can't open first link
2026-06-23 13:49:03 <vms14       > Stalevar: it's a game jam for concatenative langs
2026-06-23 13:49:22 <Stalevar    > what is game jam?
2026-06-23 13:49:22 <vms14       > you could participate with forth, or any other cat lang, or your own :D
2026-06-23 13:49:36 <Stalevar    > dog lang?
2026-06-23 13:49:47 <vms14       > Stalevar: like a contest where you make a game and see whether you are better than anyone or not
2026-06-23 13:49:57 <vms14       > no, only cats
2026-06-23 13:50:46 <Stalevar    > cheetah?
2026-06-23 13:51:05 <vms14       > i guess
2026-06-23 13:51:16 <vms14       > cheater lang
2026-06-23 13:51:38 <vms14       > that's my lang, because it cheats on everything
2026-06-23 13:54:29 <Stalevar    > https://www.freepascal.org/pic/logo.gif
2026-06-23 15:14:58 <forthBot    > Environment for Stalevar inactive, freeing...
2026-06-23 15:44:54 <Stalevar    > code incr ret ret ret ret end-code
2026-06-23 15:45:04 <Stalevar    > $50F0000003CB8 ' incr 16 + !  ok
2026-06-23 15:46:55 <Stalevar    > gforth breaks terminal if you do this (name incr is misleading. I wanted to make 1+ but didn't get it so I got down to syscall with 60 in %rax instead)
2026-06-23 17:47:57 <Stalevar    > https://github.com/jrudolph/picoforth there is also that, not only zeptoforth and mecrisp