2025-05-28 04:36:27 <deadmarshal_> I mean, in J you have to tell it on what rank of array to do operation on, but in APL it is less explicit and just does it.  idk about K
2025-05-28 08:14:47 <tpbsd       > hiall, g'day crc :) tp here
2025-05-28 08:15:22 <tpbsd       > announcing we now have a readline wrapper and command completion for mecrisp-stellaris !
2025-05-28 08:16:11 <tpbsd       > pic is at my doc site: https://sourceforge.net/projects/mecrisp-stellaris-folkdoc/files/swd2rl.pl.png
2025-05-28 08:17:16 <tpbsd       > the wrapper: https://sourceforge.net/projects/mecrisp-stellaris-folkdoc/files/swd2rl.pl
2025-05-28 09:28:49 <xentrac     > tpbsd: congratulations!
2025-05-28 09:30:55 <tpbsd       > xentrac, thanks, but Jim Morris did the design, I only applied it and wrote some support stuff :)
2025-05-28 09:30:55 <tpbsd       > at least now Mecrisp-Stellaris has a decent readline and Word completion, it's really great to use as Mecrisp-Stellaris was pretty basic
2025-05-28 09:31:36 <tpbsd       > even tho Crest wrote a excellent SWD connected terminal for it (no clock or usart issues)
2025-05-28 09:51:19 <xentrac     > that's excellent!
2025-05-28 09:53:40 <veltas      > tpbsd: Hi Terry, hope you're doing well
2025-05-28 09:54:46 <veltas      > That's great news, readline stuff is great
2025-05-28 09:55:01 <veltas      > I'll have a look tonight
2025-05-28 09:55:33 <tpbsd       > hiya veltas.I'm good thanks, I hope youre doing well also
2025-05-28 09:56:34 <tpbsd       > I'm a recent Neovim convert, and have made a couple of plugins, all Forth support stuff of course
2025-05-28 09:57:39 <tpbsd       > last time I was here, this ch was on Freenode iirc, how time flies !
2025-05-28 09:57:42 <xentrac     > can you script Neovim in Forth yet?
2025-05-28 09:58:28 <tpbsd       > xentrac, I'm a electronics guy, so I don't know, but anything is possible I guess
2025-05-28 09:59:21 <tpbsd       > Im not a programmer really, but I have to program a little to support my embedded Forth habit
2025-05-28 10:00:56 <tpbsd       > whilst I'm 100% Forth on embedded, I'm ok with Lua etc for the PC side, or even Retro :)
2025-05-28 10:17:54 <veltas      > You might be interested in https://github.com/veltas/forth-vim which is just my attempt at better syntax highlighting for Forth
2025-05-28 10:18:06 <veltas      > I imagine it's better to play with it a little for each person's needs / dialect
2025-05-28 10:18:16 <veltas      > For vim anyway
2025-05-28 10:18:27 <veltas      > I assume it's compatible with neovim?  No clue about neovim
2025-05-28 10:22:22 <tpbsd       > veltas, neovim has a excellent syntax highlighting strategy, they use a proper parser
2025-05-28 10:22:58 <tpbsd       > vim uses grep for highlighting I read
2025-05-28 10:24:48 <tpbsd       > veltas, neovim uses the "treesitter parser" for more advanced highlighting
2025-05-28 10:26:23 <tpbsd       > veltas, and then we have the so called 'language Server Protocol" servers which are just amazing and do command completion, snippets, help etc
2025-05-28 10:27:38 <tpbsd       > veltas, sadly only one Forth LSP exists and it was made for the Helix editor. LSP's were designed by Microsoft and are supposed to be compatible but I havent been able to use that one as it wont talk to me yet
2025-05-28 10:29:43 <tpbsd       > veltas, your forth-vim looks very detailed
2025-05-28 10:41:23 <tpbsd       > veltas, here is a pic I just made of a lua LSP at work in neovim, you can see how comprehensive this thing is as it's realtime, not a compiler etc :https://sourceforge.net/projects/mecrisp-stellaris-folkdoc/files/lua-lsp-2.png
2025-05-28 10:45:39 <veltas      > The ultimate syntax highlighter is the actual Forth, if you build in some 'colours' for words, because Forth has a contextual grammar after all
2025-05-28 10:45:58 <veltas      > So any parser is an approximation, short of running the code
2025-05-28 10:46:20 <tpbsd       > good point
2025-05-28 10:46:30 <veltas      > But the forth.vim I've written is just a better and less crappy approximation than what comes with vanilla vim
2025-05-28 10:46:44 <veltas      > Sounds like it might not be appropriate for neovim
2025-05-28 10:47:07 <tpbsd       > yeah the vim syntax highlighter covers every language known from a -z I think
2025-05-28 10:48:14 <tpbsd       > veltas, your highlighter wouldnt run in neovim as they use the treesitter parser  and it's not compatible with vim I'm pretty sure
2025-05-28 10:52:08 <identity    > tpbsd: iirc you can still use vim syntax highlighting packages in neovim
2025-05-28 10:53:15 <tpbsd       > identity, ah ok
2025-05-28 11:01:10 <tpbsd       > veltas, I just made this pic on my system of the neovim default highlighter the 'treesitter' parser for your viewing pleasure :) https://sourceforge.net/projects/mecrisp-stellaris-folkdoc/files/nvim-treesitter-forth.png
2025-05-28 11:05:23 <xentrac     > tpbsd: using a proper parser doesn't seem like a good idea for syntax-highlighting Forth ;-)
2025-05-28 11:06:47 <tpbsd       > xentrac, youd have to take that up with the designers, Im only a electronics guy using Forth to make my hardware work
2025-05-28 11:19:34 <tpbsd       > for those that understand this kind of stuff, here is the doc for the Python TreeSitter : https://tree-sitter.github.io/py-tree-sitter/classes/tree_sitter.Language.html
2025-05-28 11:23:09 <xentrac     > for Python it's a fine idea
2025-05-28 11:25:06 <tpbsd       > ok, I'll bite :) why doesnt it seem like a good idea for syntax-highlighting Forth when grep like methods are usually used (as in vim) ?
2025-05-28 11:26:07 <xentrac     > oh, just because the Forth parser is totally programmable
2025-05-28 11:26:43 <xentrac     > you can define your own immediate words that parse chunks out of the input using whatever delimiters they want with word
2025-05-28 11:27:16 <xentrac     > so you might as well just use regexps
2025-05-28 11:27:44 <xentrac     > unless you're going to trace the execution of a Forth system as it compiles your source code ;)
2025-05-28 11:28:25 <tpbsd       > I only do that when my Forth source is causing exceptions etc
2025-05-28 11:28:34 <tpbsd       > which is probably a lot
2025-05-28 11:28:43 <xentrac     > I mean, you could do it automatically in order to get syntax highlighting information
2025-05-28 11:29:55 <tpbsd       > xentrac, youre talking about a pc hosted Forth I assume ? I only use tiny mcus for forth, the PC has to supply the terminal and intelligence in my case
2025-05-28 11:30:55 <tpbsd       > I realise that this ch is all about PC hosted forths, thats why I havent visited in a few years, your Forth and my Forth are very different
2025-05-28 11:31:38 <tpbsd       > mine has like 8k of ram and 64kb of flash, it can only perform a dedicated and small device function, always standalone
2025-05-28 11:31:57 <tpbsd       > no usb etc
2025-05-28 11:34:38 <xentrac     > not necessarily, no
2025-05-28 11:34:50 <xentrac     > I think MCU-hosted Forths are a lot more interesting actually
2025-05-28 11:35:46 <xentrac     > but what I mean is that if you define a word : vec3 create , , , ; and then you use it like 3 4 12 vec3 foo what you want to know from the syntax highlighting is that foo is being defined there
2025-05-28 11:36:31 <xentrac     > maybe it would even be useful to know that it's defined as a vec3
2025-05-28 11:37:05 <xentrac     > to figure that out you need to trace the compilation process and observe that the characters "foo" are being consumed by create (which is being called from vec3)
2025-05-28 11:38:16 <xentrac     > there are lots of ways you could conceivably do that, including doing the compilation process on the PC and just downloading the compiled dictionary to the board, tracing the compilation over OCD or similar, or running the compilation in an emulator
2025-05-28 11:38:26 <xentrac     > but that kind of analysis is fundamentally beyond tree-sitter's capabilities
2025-05-28 11:38:43 <tpbsd       > is it beyond a LSP's capability ?
2025-05-28 11:39:40 <tpbsd       > I do know that this level of software is definitely beyond *my* capability :)
2025-05-28 11:40:33 <xentrac     > no, an LSP server can totally provide the editor with that information; kind of the whole point of LSP is really to enable editors to talk to compilers
2025-05-28 11:42:19 <tpbsd       > xentrac, so perhaps use a LSP to provide the highlighting then ?
2025-05-28 11:42:56 <tpbsd       > sadly, there is only one and it's basic I believe, and I dont know how to install it in neovim
2025-05-28 11:43:18 <tpbsd       > as it was made for a different editor, one with no plugins
2025-05-28 11:44:06 <tpbsd       > it seems plenty of programmers dont want to mess about with a editor to make it useful, but Im ok with that, not being a real programmer myself
2025-05-28 11:44:22 <xentrac     > the LSP probably needs to encapsulate the Forth system you're actually using
2025-05-28 11:44:31 <tpbsd       > I must try that other editor sometime
2025-05-28 11:45:04 <tpbsd       > yeah, because 'once you know one Forth ... then you only know *one* Forth' ;-)
2025-05-28 11:46:13 <xentrac     > well, I mean, if you have something in your compilation early on that errors out in the wrong Forth, that kind of thing
2025-05-28 11:46:21 <tpbsd       > I mean, Im a simple noob class Forth user myself, but I wont use anything else for embedded except mecrisp_Across or assembly now and again
2025-05-28 11:46:43 <tpbsd       > yeah, idealy it should be able to adapt easily
2025-05-28 11:46:56 <xentrac     > yeah
2025-05-28 11:49:06 <tpbsd       > xentrac, do you know the Gema text replacement prog ?
2025-05-28 11:50:52 <tpbsd       > Ive used it to build a system where I use labels for mcu register fields, and the system replaces them with their register absolute address on the way to the mcu. The source is untouched
2025-05-28 11:52:16 <xentrac     > is it like the C preprocessor?
2025-05-28 11:53:25 <tpbsd       > I really dont know, I had the impression that  C preprocessor is very complex
2025-05-28 11:55:01 <tpbsd       > the main thing with Gema is that I feed it huge text files of text replacement config which I generate from the MCU manufacturers XML System View Description file using XSLT
2025-05-28 11:55:35 <tpbsd       > so the whole process is automatic, it just needs a SVD for the specific MCU
2025-05-28 11:56:11 <tpbsd       > it's been working perfectly for a couple of years now, so I will release it as soon as I can tidy it all up etc
2025-05-28 11:56:39 <tpbsd       > it basically removes all 'magic numbers' from the Forth code
2025-05-28 11:56:58 <tpbsd       > replacing them with CMSIS-SVD nomenclature
2025-05-28 12:00:50 <xentrac     > I'm not familiar with CMSIS-SVD at all
2025-05-28 12:01:20 <tpbsd       > thats cool, what I do is very obscure and niche
2025-05-28 12:01:27 <xentrac     > the C preprocessor is commonly used for that purpose though
2025-05-28 12:01:33 <tpbsd       > ahh
2025-05-28 12:01:42 <xentrac     > it does have some tricky aspects
2025-05-28 12:01:53 <tpbsd       > I found Gema fairly easy to grok, took me a few months tho
2025-05-28 12:01:59 <tpbsd       > I bet it does
2025-05-28 12:02:48 <tpbsd       > gema is a general purpose text processing utility based on the concept of pattern matching. It reads an input file and copies it to an output file transforming the data as specified by the patterns defined by the user.
2025-05-28 12:03:03 <xentrac     > like for example
2025-05-28 12:03:09 <xentrac     > #define ARDUINO_MODEL_USB_PID 0x0038
2025-05-28 12:03:21 <tpbsd       > if your interested in it : https://sourceforge.net/projects/gema/
2025-05-28 12:03:28 <xentrac     > causes the identifier ARDUINO_MODEL_USB_PID to be replaced thereafter with the hex number 0x0038
2025-05-28 12:03:46 <tpbsd       > thats easy to see
2025-05-28 12:04:19 <xentrac     > which is handy for things like register addresses and bit masks
2025-05-28 12:05:09 <tpbsd       > sure
2025-05-28 12:05:17 <tpbsd       > btw: The CMSIS System View Description format (CMSIS-SVD) formalizes the description of the system contained in Arm Cortex-M processor-based micro controllers, in particular, the memory mapped registers of peripherals. The detail contained in system view descriptions is comparable to the data in device reference manuals.
2025-05-28 12:05:19 <xentrac     > thanks for the reference to Gema!  I'll take a look; it sounds like it could be interesting
2025-05-28 12:06:09 <tpbsd       > I really found it incredibly useful and after I broke my brain on it and gave up all hope, everything was much easier ;-)
2025-05-28 12:06:39 <xentrac     > haha
2025-05-28 12:06:57 <tpbsd       > one problem I have with parsers is when they generate no output because of my brain dead patterns
2025-05-28 12:07:18 <xentrac     > Unix has a lot of general purpose text processing utilities based on the concept of pattern matching
2025-05-28 12:07:32 <xentrac     > I think the Unix people designed the C preprocessor after burning their fingers a lot on some of the older ones
2025-05-28 12:07:33 <tpbsd       > so once I learned to not take being totally ignored personally, that got easuer also ;-)
2025-05-28 12:07:40 <xentrac     > haha
2025-05-28 12:08:25 <tpbsd       > however if I had to master REGEX or die, I'd say 'just kill me now!'
2025-05-28 12:09:51 <tpbsd       > I was going to look into the newish Java parser as it seems to inform whats happening instead of the 'cold shoulder treatment' but I couldnt bring myself to inflict Java on my users
2025-05-28 12:10:52 <xentrac     > I've been thinking about a new text pattern matching system I call the Monkey's Paw
2025-05-28 12:10:52 <xentrac     > with some improvements over regexps
2025-05-28 12:11:10 <tpbsd       > so Ive stuck with Gema. Using Gema is like being made into a Necromonger (Croicles of Riddick) the pain goes away after a while
2025-05-28 12:11:32 <tpbsd       > chronicles
2025-05-28 12:11:52 <tpbsd       > xentrac, that sounds really complex
2025-05-28 12:12:53 <tpbsd       > xentrac, have a look at Gema first, try it out perhaps
2025-05-28 12:13:51 <tpbsd       > xentrac, gema can also do multi-line replacements !
2025-05-28 12:14:06 <xentrac     > so for example a Monkey's Paw pattern is <s: <t: <v: [a-z]+> | <n: [0-9]+> | "(" <s> ")"> (<op: "+" | "-"> <t>)*>
2025-05-28 12:14:44 <xentrac     > which matches integer arithmetic expressions of sums and differences with properly balanced parentheses
2025-05-28 12:14:48 <xentrac     > and variables
2025-05-28 12:15:16 <tpbsd       > is that the matching config only ?
2025-05-28 12:15:25 <xentrac     > yeah
2025-05-28 12:15:54 <tpbsd       > Gema has a lhs and a rhs config
2025-05-28 12:16:07 <xentrac     > you can include "semantic actions" too but the example gets to be longer than one line :)
2025-05-28 12:16:20 <tpbsd       > whats on the left gets replaced by whats on the right, a '=' separates them
2025-05-28 12:16:41 <xentrac     > yeah
2025-05-28 12:17:18 <cleobuline  > 7 hours to  calculate 1 859433  <<  1 -  PRIME? .
2025-05-28 12:25:09 <cleobuline  > forthBot:  MILLI 1 859433  <<  1 -  PRIME? . MILLI SWAP - .  ."  1 859433  <<  1 -  est premier "
2025-05-28 12:25:20 <tpbsd       > hahah, oh wow. I'm running a Mecrisp-Stellaris SWDCOM terminal in a python wrapper in the Neovim terminal and it's working!!!
2025-05-28 12:25:39 <tpbsd       > Im talking to Mecrisp-Stellaris Forth in the mcu flash
2025-05-28 12:44:20 <tpbsd       > https://sourceforge.net/projects/mecrisp-stellaris-folkdoc/files/neovim-terminal-split.png
2025-05-28 14:36:34 <cleobuline  > forthBot:  WORDS
2025-05-28 16:22:45 <MrMobius    > speaking of PC vs MCU forth, ive wondered about having a tiny interpreter on the MCU and interpreted code on an SPI EEPROM
2025-05-28 16:23:32 <MrMobius    > so it could poduce it's own optimized code like mecrisp but only write the relevant part to it's own flash
2025-05-28 16:27:14 <xentrac     > tphey, that's fantastic!
2025-05-28 16:27:19 <xentrac     > oops
2025-05-28 16:27:22 <xentrac     > too late!
2025-05-28 16:27:39 <xentrac     > MrMobius: like the BASIC Stamp?
2025-05-28 16:28:05 <MrMobius    > could be. I dont know how the BASIC Stamp works
2025-05-28 16:28:14 <MrMobius    > I did get one for free once but never did anything with it
2025-05-28 16:29:28 <MrMobius    > wikipedia seems to say the interpreter is on the MCU
2025-05-28 16:29:42 <xentrac     > http://www.mcmanis.com/chuck/robotics/stamp-decode.html Chuck McManis’s teardown of the Parallax BASIC Stamp #microcontrollers firmware, with its variable-bit-length instructions and operands.
2025-05-28 16:31:22 <MrMobius    > "BASIC programs are compiled into tokens on the PC and then serially downloaded into the EEPROM on the stamp board."
2025-05-28 16:32:32 <MrMobius    > I was thinking of bytecode on the EEPROM that implements a Forth prompt and compiler. Compiled programs, however, are written to the MCU's flash as optimized machine code so you dont need the EEPROM after youre done programming
2025-05-28 16:33:50 <MrMobius    > so instead of a full forth taking up most of the room on the MCU, it's just a very small interpreter and the rest is free for machine code
2025-05-28 16:37:57 <xentrac     > I see
2025-05-28 16:38:20 <xentrac     > it might depend a bit on how small of a microcontroller you're talking about
2025-05-28 16:40:29 <xentrac     > the PY32F002B is 8.51¢, 24 kibibytes flash, 3 kibibytes SRAM
2025-05-28 16:41:07 <xentrac     > I think it's fairly straightforward to get a Forth prompt and compiler into 4 kibibytes, and you might be able to make it in 2
2025-05-28 16:48:03 <MrMobius    > Mecrisp needs about 20K IIRC to produce optimized machine code
2025-05-28 16:48:17 <MrMobius    > it would be nice to use it with msp430s which have max 16k flash in dip
2025-05-28 16:49:00 <MrMobius    > Im sure you can interpret the EEPROM code in 1K or less
2025-05-28 16:49:21 <MrMobius    > and probably like 100 bytes if you copy little snippets of machine code into RAM
2025-05-28 16:51:32 <xentrac     > yeah, I meant a threaded-code compiler, not a native-code compiler
2025-05-28 16:52:00 <MrMobius    > right. you could do that but then no need for the EEPROM. I was just thinking of a way to get optimized machine code without a tethered board
2025-05-28 16:54:05 <xentrac     > it's an interesting question\
2025-05-28 16:55:20 <MrMobius    > ya interesting especially if youre trying not to use much RAM
2025-05-28 16:55:43 <MrMobius    > since you would hopefully write things to a scratch space in EEPROM as you optimize but not change them after writing
2025-05-28 16:56:04 <xentrac     > the freedom Forth affords the programmer makes an optimizing compiler more difficult, I think.  the compiler really wants to know the stack effects of the words it's compiling
2025-05-28 16:58:14 <MrMobius    > exactly. that's how mecrisp does it
2025-05-28 17:00:49 <xentrac     > I mean that's step 1
2025-05-28 17:01:09 <xentrac     > it's hard to do much if you can't do that
2025-05-28 17:02:14 <xentrac     > I know this is heretical but maybe Lisp would be a better fit if you want an optimizing native code compiler running on the microcontroller itself
2025-05-28 19:03:14 <MrMobius    > bu then id have to write lisp code :)
2025-05-28 19:27:15 <skvery      > Mecrisp is very fast.
2025-05-28 19:28:34 <skvery      > The Forth library is pre-assembled, so it is jyst a list of subroutine calls in the end.
2025-05-28 19:42:48 <forthBot    > 1
2025-05-28 19:42:48 <forthBot    > 26258789
2025-05-28 19:42:49 <forthBot    >  1 859433  <<  1 -  est premier
2025-05-28 19:42:49 <forthBot    > USERNAME .S . + - * / MOD DUP DROP SWAP OVER ROT >R R> R@ R@UL R!UL = < > AND OR NOT XOR & | ^ ~ << >> CR EMIT VARIABLE @ ! +! DO LOOP I WORDS LOAD CREATE ALLOT ." CLOCK BEGIN WHILE REPEAT AGAIN SQRT UNLOOP +LOOP PICK CLEAR-STACK PRINT NUM-TO-BIN PRIME? FORGET STRING S" "S 2DROP IMAGE TEMP-IMAGE CLEAR-STRINGS DELAY EXIT CONSTANT MICRO MILLI ROLL DEPTH APPEND
2025-05-28 20:28:32 <cleobuline  > forthBot:  26258789 3600000 / .
2025-05-28 20:28:32 <forthBot    > 7
2025-05-28 20:40:38 <vms14       > I have the word SEE
2025-05-28 20:40:40 <vms14       > https://i.imgur.com/c2T2nry.png
2025-05-28 20:40:50 <vms14       > but it cannot show the code of a colon definition xd
2025-05-28 20:40:55 <vms14       > only primitives
2025-05-28 20:41:53 <vms14       > drop is actually the get function
2025-05-28 20:43:28 <vms14       > I could make it show colon definitions, but they are actually that subroutine shown and I'm not going to add metadata for now nor expose that @code array
2025-05-28 20:52:29 <MrMobius    > hmm, I wonder if having some buffer before and after the stack and checking for under/overflow every so often is faster than a circular stack
2025-05-28 20:53:42 <MrMobius    > since if you check every 5th instruction, you only need enough buffer on either side for the worst case that 5 instructions can do
2025-05-28 20:54:47 <MrMobius    > it just wouldnt catch briefly going out of bounds then back in but circular stacks dont either
2025-05-28 20:58:13 <MrMobius    > maybe ORing the most extreme address each primitive uses and ANDing it with another. highest bits will show out of bounds if stack is aligned
2025-05-28 22:28:32 <forthBot    > Environment for cleobuline inactive, freeing...
2025-05-28 22:40:30 <tpbsd       > MrMobius, I took your advice and joined liberia ##forth, I caught up with some people I know and had a good time, theyre a great bunch!
2025-05-28 22:41:16 <tpbsd       > oops wrong ch!
2025-05-28 22:41:56 <tpbsd       > Im still half asleep :)
2025-05-28 22:50:32 <MrMobius    > tpbsd: welcome! I saw your messages before but you had left
2025-05-28 22:51:56 <tpbsd       > MrMobius, I'm on cellphone Internet and I disconnect it at night when I sleep to save quota
2025-05-28 22:52:13 <MrMobius    > ah makes sense. no more starlink?
2025-05-28 22:52:31 <tpbsd       > I can leave it connected when I resub to Starlink
2025-05-28 22:53:44 <tpbsd       > no, not atm, I blew my budget building a pc especially for AI and now have to tighten my belt for a while. Starlink is $139AUD a month, cell is $39/month
2025-05-28 22:54:45 <tpbsd       > MrMobius, I really like starlink, it's a step up from cell, never any bandwidth issues and it's unlimited
2025-05-28 22:55:59 <tpbsd       > MrMobius, my starlink terminal is mounted outside on the wall and waits patiently for resub!
2025-05-28 22:56:42 <tpbsd       > I can do it in 2 minutes and Im on as soon as the fee is paid, it's excellent
2025-05-28 23:04:39 <xentrac     > tpbsd: hey, that's fantastic!
2025-05-28 23:04:48 <xentrac     > (the SWDCOM terminal)
2025-05-28 23:05:11 <tpbsd       > xentrac, it is awesome and 100% stable
2025-05-28 23:06:15 <xentrac     > MrMobius and I were speculating after you left about on-device optimizing native-code compilation
2025-05-28 23:06:33 <tpbsd       > xentrac, but there is one small problem (I think) it looks like neovim :terminal lacks USB 'backpressure', so there is no flow control and the Forth chip chokes when compiling
2025-05-28 23:07:14 <tpbsd       > xentrac, so I'll be trying some workarounds today
2025-05-28 23:07:21 <xentrac     > no flow control?
2025-05-28 23:07:28 <xentrac     > maybe flow control is just misconfigured
2025-05-28 23:07:47 <tpbsd       > looks like as it chokes when uploading large source files
2025-05-28 23:08:12 <tpbsd       > xentrac, there is no mention of flow control in their docs
2025-05-28 23:08:39 <xentrac     > hmph!
2025-05-28 23:09:01 <tpbsd       > I'm guessing they havent faced that issue as yet
2025-05-28 23:09:17 <tpbsd       > expecially USB 'backpressure'
2025-05-28 23:09:52 <tpbsd       > Im using SWD (JTAG) for the terminal, not a on chip usart
2025-05-28 23:10:16 <tpbsd       > and it's very fast, about equiv to 1Mbaud +
2025-05-28 23:11:38 <tpbsd       > in fact it's so fast uploading that one cant read the terminal as it uploads, so I had to add a beep on error so if I hear a beep I know the upload failed
2025-05-28 23:12:07 <tpbsd       > and Ive the slow one!, Crest runs his with a STLINK3 at 25Mhz
2025-05-28 23:13:00 <tpbsd       > but it needs a STM32H7xx series chip at 280MHz to keep up while compiling
2025-05-28 23:15:15 <xentrac     > yeah, I'm just wondering if SWD has some kind of flow control.  or maybe multiple forms of flow control
2025-05-28 23:15:35 <xentrac     > or SWDCOM rather
2025-05-28 23:16:23 <xentrac     > an extra diode and capacitor isn't that many extra passives
2025-05-28 23:20:28 <tpbsd       > xentrac, SWDCOM has USB 'backpressure' flow control
2025-05-28 23:21:38 <tpbsd       > xentrac, which it needs because a USB/SWD interface is always used, usually one of the $2 Chinese ones
2025-05-28 23:22:17 <tpbsd       > or a STM Stlink V3 which only STM make as far as I know
2025-05-28 23:22:59 <tpbsd       > I have 4 of them but havent tried them yet because they use the minature connectors
2025-05-28 23:23:53 <tpbsd       > besides, I always develop on the STM32F051 Cortex-M0 which has a max clock of 48MHz
2025-05-28 23:24:31 <tpbsd       > so a STlink v2 is all I need for max upload speed with that chip
2025-05-28 23:28:57 <xentrac     > the STM32 and clones are pretty astounding little machines
2025-05-28 23:30:20 <tpbsd       > xentrac, Im 70 years old and have been doing embedded since the 70's and Ive never seen such cheap and amazing mcu's
2025-05-28 23:37:49 <xentrac     > there are cheaper ones out there than the STM32
2025-05-28 23:38:33 <xentrac     > AFAICT the PY32F002 is not an STM32 clone, but it's still a 48MHz ARM, and it's 9.01¢
2025-05-28 23:39:18 <tpbsd       > xentrac, id happily pay $40 for a single mcu, Im from the 70's !
2025-05-28 23:39:43 <tpbsd       > xentrac, the price of a mcu is utterly irrelevant to me
2025-05-28 23:39:55 <tpbsd       > it's my time thats limited and valuable
2025-05-28 23:40:58 <tpbsd       > xentrac, have you seen my Mecrisp-Stellaris documentation site ? https://mecrisp-stellaris-folkdoc.sourceforge.io/
2025-05-28 23:41:27 <tpbsd       > it goes back to 2014 when I started learning Forth and using STM mcus
2025-05-28 23:41:27 <xentrac     > surely you can think of something useful to do with 500 MCUs, especially if you can get them reflow-soldered with a pick-and-place machine
2025-05-28 23:41:58 <xentrac     > it's a very nice site with awesome projects
2025-05-28 23:42:24 <tpbsd       > xentrac, well it's a hobby now as Im long retired, and living in Australia means it's hard to get reliable supply, so I hoard instead
2025-05-28 23:42:38 <xentrac     > that reminds me, is it feasible to detect CO optically?
2025-05-28 23:42:50 <tpbsd       > sure
2025-05-28 23:43:02 <tpbsd       > it's complex tho
2025-05-28 23:43:07 <xentrac     > what wavelengths do you use?  do you have to do it photoacoustically?
2025-05-28 23:43:15 <tpbsd       > I used to calibrate CO machines
2025-05-28 23:43:24 <tpbsd       > I know then very well
2025-05-28 23:43:44 <xentrac     > cheap MCUs favor hoarding because you can buy 1000 for US$90
2025-05-28 23:43:45 <tpbsd       > you use infra red, basically a very bull filament
2025-05-28 23:43:52 <xentrac     > dull?
2025-05-28 23:43:58 <tpbsd       > oops yeah
2025-05-28 23:44:06 <tpbsd       > chery red
2025-05-28 23:44:23 <xentrac     > and then detect a narrow absorption line?
2025-05-28 23:44:35 <tpbsd       > you need a chopper wheel, a thermo sensor like they use in alarms
2025-05-28 23:44:49 <xentrac     > like a PIR proximity sensor?
2025-05-28 23:45:02 <tpbsd       > yeah, a filter is used, it's made of some salt crystals, you cant see thru it
2025-05-28 23:45:11 <tpbsd       > exactly like one of those
2025-05-28 23:45:21 <xentrac     > hmm, I wonder which salt
2025-05-28 23:45:53 <tpbsd       > I never knew that, the machines were made in Australia by 'repco', but there are american and japanese ones
2025-05-28 23:45:57 <xentrac     > it's easy to make CO for qualitative calibration testing, I think, with charcoal and damp brick
2025-05-28 23:46:07 <xentrac     > (burning charcoal)
2025-05-28 23:46:22 <tpbsd       > or a car engine with no cat
2025-05-28 23:46:23 <tpbsd       > ;-)
2025-05-28 23:46:26 <xentrac     > what's the difficulty with reliable supply in Australia?
2025-05-28 23:46:33 <xentrac     > car engines are a lot more expensive than charcoal
2025-05-28 23:46:39 <tpbsd       > distance and suppliers
2025-05-28 23:46:45 <tpbsd       > true
2025-05-28 23:47:08 <tpbsd       > we had very expensive bottles of calibrated CO, parts per million
2025-05-28 23:47:34 <xentrac     > haha
2025-05-28 23:47:40 <tpbsd       > such a machine is very reliable and accurate for CO measurement
2025-05-28 23:48:04 <xentrac     > I'd be happy to just have an alarm if the CO level in my house went over 1000ppm, ±50%
2025-05-28 23:48:25 <tpbsd       > you also have the 'figaro' line of CO sensors just used for detection of CO in underground carparks etc
2025-05-28 23:48:51 <tpbsd       > ahh, then you dont need a $4000 gas analyser!
2025-05-28 23:49:10 <tpbsd       > some of the 'figaro' sensors would do
2025-05-28 23:49:23 <tpbsd       > I made up a heap of such sensors for a carpark once
2025-05-28 23:49:57 <xentrac     > yeah, something like that
2025-05-28 23:50:09 <tpbsd       > Im guessing you could make up something with the right led and a pump etc
2025-05-28 23:50:53 <xentrac     > evidently they aren't available in my country: https://listado.mercadolibre.com.ar/figaro-sensor#D[A:figaro%20sensor]
2025-05-28 23:51:12 <tpbsd       > talk to Matthias Koch (the mecrisp* dev) as hes big on ramen sensors and a chemist
2025-05-28 23:51:20 <xentrac     > oh cool!
2025-05-28 23:51:35 <tpbsd       > he would be the very best source there
2025-05-28 23:51:37 <xentrac     > I feel like a chopper wheel, nichrome filament, and PIR sensor should be quite DIYable
2025-05-28 23:51:44 <tpbsd       > me too
2025-05-28 23:51:48 <xentrac     > but exotic salts might not be
2025-05-28 23:51:50 <tpbsd       > and a air pump
2025-05-28 23:52:19 <tpbsd       > yeah the world is nuts
2025-05-28 23:52:23 <anthk_      > where?
2025-05-28 23:52:23 <xentrac     > sodium hydroxide, MEK, HCl
2025-05-28 23:52:28 <xentrac     > Argentina
2025-05-28 23:52:55 <anthk_      > ah, sodium hydrox might be legal to fetch in Spain
2025-05-28 23:53:23 <anthk_      > that's just a common agent to produce soap
2025-05-28 23:53:28 <tpbsd       > some gang characters fought with machetes in Melbourne recently, so now theyre banning machetes
2025-05-28 23:53:56 <xentrac     > hahaha
2025-05-28 23:53:58 <anthk_      > caustic soda it's uber common
2025-05-28 23:54:06 <tpbsd       > theyre not banning gang characters ...
2025-05-28 23:54:19 <xentrac     > back to medieval times: only the nobility are allowed to carry swordsw
2025-05-28 23:54:23 <anthk_      > every granma from a village knows how to use that
2025-05-28 23:54:27 <anthk_      > xentrac: not true
2025-05-28 23:54:39 <anthk_      > tons of people had small swords and knife
2025-05-28 23:54:47 <anthk_      > knives/daggers
2025-05-28 23:54:57 <xentrac     > oh, sure, there were places where peasants were permitted swords in medieval times
2025-05-28 23:54:59 <tpbsd       > anthk_, xentrac is correct in general
2025-05-28 23:55:04 <xentrac     > but there were lots of places where they weren't
2025-05-28 23:55:22 <tpbsd       > look at japan ? only samurai were allowed to carry swords
2025-05-28 23:55:27 <anthk_      > still, peasants got several ways to use weapons
2025-05-28 23:55:35 <tpbsd       > anyone else would be killed if they did
2025-05-28 23:55:40 <xentrac     > with a sword
2025-05-28 23:55:42 <xentrac     > now the samurai are called "police"
2025-05-28 23:55:43 <anthk_      > clubs, field farming tools, and so on
2025-05-28 23:55:49 <tpbsd       > yeah, hidden daggers etc
2025-05-28 23:55:50 <anthk_      > heh, that reminds me of Dragon Ball
2025-05-28 23:56:00 <anthk_      > goku it's the peasant with a bo-ken (stick)
2025-05-28 23:56:07 <anthk_      > against vegeta, some space samurai
2025-05-28 23:56:16 <veltas      > deep
2025-05-28 23:56:27 <tpbsd       > hey veltas :)
2025-05-28 23:56:29 <xentrac     > yeah, Musashi supposedly won lots of duels armed only with boken
2025-05-28 23:56:36 <veltas      > Hi Terry
2025-05-28 23:56:42 <anthk_      > well, that's true, the Saiyan armor from DB it's a copycat from the Samurai ones
2025-05-28 23:57:03 <veltas      > This is a very good example of Forth on MCU vs C https://mecrisp-stellaris-folkdoc.sourceforge.io/peripherals-c-vs-svd2forth.html
2025-05-28 23:57:20 <tpbsd       > it seems that martial arts were developed because peasants were not allowed swords etc
2025-05-28 23:57:50 <anthk_      > indeed
2025-05-28 23:57:54 <anthk_      > Karate
2025-05-28 23:57:59 <anthk_      > tons of poses are from archery
2025-05-28 23:58:00 <xentrac     > technically I think samurai were strictly modern, not medieval
2025-05-28 23:58:19 <xentrac     > I'm not sure the Shaolin monks were technically peasants
2025-05-28 23:58:39 <anthk_      > church
2025-05-28 23:58:46 <tpbsd       > veltas, Im glad you like it :)
2025-05-28 23:58:53 <anthk_      > another branch of feudalism
2025-05-28 23:59:18 <anthk_      > brb, switch back from emwm to cw
2025-05-28 23:59:20 <anthk_      > cwm
2025-05-28 23:59:37 <tpbsd       > veltas, Ive moved on since then with my 'Plang2' system for Mecrisp-Stellaris
2025-05-28 23:59:49 <xentrac     > I'm not sure exactly what the historical development of gong fu was