2024-05-27 15:19:09 <veltas      > Yeah when you interpret a block the block's buffer is the input buffer, when you interpret at the 'interpreter' a line is the buffer
2024-05-27 15:20:09 <veltas      > So the next line isn't loaded when you parse with CREATE , it should just fail if there's nothing on same line
2024-05-27 15:21:40 <zelgomer    > why would't it just block inside of the "word" called from create until you enter the next line?
2024-05-27 15:22:36 <veltas      > There are forths that do that
2024-05-27 15:22:47 <veltas      > Probably only legacy reasons why it isn't that way in the standard
2024-05-27 15:23:06 <veltas      > Classic forths don't do this I think, probably because there was no grand unified 'REFILL' thing
2024-05-27 15:23:24 <zelgomer    > mine does, which is why i'm surprised they wouldn't all do it
2024-05-27 15:24:13 <zelgomer    > if your parsing is done at the callee level then it just happens naturally
2024-05-27 15:28:54 <veltas      > All you have to do with a standard-style forth is make e.g. PARSE use REFILL, and it makes QUIT simpler as well
2024-05-27 15:29:02 <veltas      > It's just not standard for some reason
2024-05-27 15:29:11 <veltas      > I think it's historical
2024-05-27 15:30:41 <veltas      > There's even exceptions for words like ( to make them parse more than one line in files
2024-05-27 15:31:30 <zelgomer    > that sounds kludgy
2024-05-27 15:31:50 <veltas      > The standard is full of kludge
2024-05-27 15:31:55 <veltas      > That's why I don't like it
2024-05-27 15:41:40 <veltas      > It is a monument of compromise
2024-05-27 16:25:33 <veltas      > I guess one reason is that e.g. 1 WORD is expected to parse the rest of the line
2024-05-27 16:25:59 <veltas      > Or CHAR ^ WORD for editor example
2024-05-27 16:35:25 <olle        > Are there ways to indent forth code to get around the hard-to-read post-fix notation?
2024-05-27 16:36:29 <veltas      > The main way to visually group words like you would with e.g. brackets is using extra spaces
2024-05-27 16:36:46 <veltas      > e.g. 2 or 3 spaces to show some words are 'grouped'
2024-05-27 16:37:23 <veltas      > The other way is factoring, as usually any group of words can just be cut and pasted into a word def above the current def
2024-05-27 16:38:38 <veltas      > You'll see in Starting Forth there is a style to group words in IF..THEN or BEGIN..UNTIL etc constructs that might be a little unfamiliar
2024-05-27 16:38:51 <veltas      > Some people use groupings/indents more like C or other code
2024-05-27 16:39:26 <veltas      > And if you've got a quotation-heavy forth like retro I could imagine using quotations for visual grouping
2024-05-27 16:39:49 <veltas      > And KipIngram's forth has a balanced stack grouping feature that uses brackets or parens or something as well
2024-05-27 16:40:04 <veltas      > Brackets or braces (?)
2024-05-27 16:40:49 <olle        > what about mathematical calculations? e.g. 100 1 "purchase_price" "selling_price" / - * 2 round
2024-05-27 16:40:59 <olle        > that's margin of profit per article
2024-05-27 16:41:17 <olle        > idk, maybe just new lines
2024-05-27 16:41:44 <olle        > And indentation
2024-05-27 16:41:54 <veltas      > 1  PURCHASE SELLING /  -  100 *  2 ROUND
2024-05-27 16:42:05 <olle        > :d
2024-05-27 16:42:15 <olle        > Maybe maybe, not too bad :)
2024-05-27 16:42:25 <veltas      > So see I group PURCHASE SELLING / together and then combine ops where possible to make the postfix less difficult
2024-05-27 16:42:59 <veltas      > Well you can see what I've done anyway.....
2024-05-27 16:43:10 <veltas      > I explain the answer worse than the answer itself
2024-05-27 16:43:58 <veltas      > olle: I tend to use lines to track stack contents, if it's 'trivial' or 'mostly-algebra' I'll put it all on one line
2024-05-27 16:44:22 <veltas      > If the stack contents is ever different by the end of the line I'll comment it in a new stack comment
2024-05-27 16:44:45 <veltas      > Unless it's the 'end' of a construct / word because then it's implied by outer comment or the word's stack comment
2024-05-27 16:45:10 <veltas      > And I don't comment return stack if I can help it, that should mostly be used like a push/pop situation which should be easy to remember
2024-05-27 16:45:34 <veltas      > That's what I've settled on in my quite verbose forth style, in my less verbose style I use more variables/words if it gets too hard
2024-05-27 16:46:12 <veltas      > If you ever get stuck just post it here, there are people here who are happy to refactor and format what you've written and will do a 10x better job than me
2024-05-27 16:46:31 <olle        > Oh this is just for my DSL project :)
2024-05-27 16:46:43 <olle        > Wondered if Forth-like DSL can be made more readable or not
2024-05-27 16:46:51 <veltas      > Forth's postfix is more readable than other langs a lot of time
2024-05-27 16:47:47 <veltas      > e.g. what's more readable for specifying your TTY:  cols 80 rows 40 baud 115200   or   80 COLS  40 ROWS  115200 BAUD
2024-05-27 16:48:40 <veltas      > You just have to lean into the postfix and understand its strengths
2024-05-27 16:49:26 <veltas      > This is why people bang on about Forth being a DSL-generator, because you can define any 'word' to do literally anything you want, and it's relatively easy to support all sorts of numbers and strings in the parsing too
2024-05-27 16:52:28 <olle        > In some cases sure
2024-05-27 16:52:48 <olle        > But my example above kills it :d
2024-05-27 16:54:14 <olle        > S-expr is more like already existing notation
2024-05-27 16:57:21 <olle        > Sad to say, I like my little Forth-inside-PHP engine ^^
2024-05-27 16:58:02 <veltas      > S-expr isn't much better than postfix when you're used to it
2024-05-27 16:58:14 <veltas      > Both are way harder to work with for non-tech people that are used to normal algebra
2024-05-27 16:58:25 <olle        > Sure, but I need semi-tech people to get used to it ;)
2024-05-27 16:58:28 <olle        > So need a lower threshold
2024-05-27 16:59:25 <olle        > "way" harder? Depends on which changes they are expected to make in a small script.
2024-05-27 17:00:28 <veltas      > Yeah
2024-05-27 17:00:45 <veltas      > I'm working on an example for you that you might like
2024-05-27 17:00:57 <veltas      > friendlier syntax closer to S-expr, simple to implement in forth
2024-05-27 17:08:56 <olle        > Yea? :)
2024-05-27 17:11:34 <zelgomer    > : compliment  1 swap - ;  : %  100 * 2 round ;  purchase selling / compliment %
2024-05-27 17:12:14 <veltas      > olle: https://termbin.com/jyyq
2024-05-27 17:12:42 <veltas      > Use a separate stack if you want return stack free for other things
2024-05-27 17:12:46 <olle        > Oh is this S-expr in forth? :D
2024-05-27 17:13:11 <olle        > zelgomer: oh not bad
2024-05-27 17:13:45 <olle        > Might need some stack juggling for order of args, but in else nice
2024-05-27 17:13:50 <olle        > -in
2024-05-27 17:14:03 <olle        > Oh you did swap
2024-05-27 17:14:05 <olle        > Sorry
2024-05-27 17:16:30 <veltas      > Anyway so you can have essentially s-exprs in forth with 3 lines of code
2024-05-27 17:19:17 <olle        > Kewl :)
2024-05-27 17:19:32 <veltas      > I would put them in a different stack tho
2024-05-27 17:19:39 <veltas      > To allow return stack usage
2024-05-27 17:19:51 <veltas      > But for your purposes that's unnecessary anyway
2024-05-27 17:20:12 <veltas      > Well actually maybe not, if they want do loops
2024-05-27 17:24:20 <olle        > I don't want to waste your time, but I actually do not have access to a full forth system, only what I can write in ~50 lines of PHP
2024-05-27 17:27:18 <olle        > Still, saved your notes :)
2024-05-27 17:35:06 <veltas      > https://termbin.com/cp8s
2024-05-27 17:35:11 <veltas      > With separate stack
2024-05-27 17:35:33 <veltas      > That's fine, I don't have any expectations
2024-05-27 17:35:57 <veltas      > The notes are useful outside of what you're doing, probably
2024-05-27 17:40:25 <veltas      > Okay one more https://termbin.com/foza
2024-05-27 17:40:28 <veltas      > Done
2024-05-27 17:40:42 <veltas      > I'm sure this has been done many times before now
2024-05-27 17:48:42 <veltas      > https://termbin.com/l56i
2024-05-27 17:49:22 <veltas      > There is a satisfaction to refining things in Forth
2024-05-27 19:30:50 <KipIngram   > olle: What veltas is referring to, I think, is my stack frae setup.  I reserved a register I call the frae pointer.  The word { saves FP on the return stack and copies SP to FP.  The word } restores SP from FP, recovers old FP from the return stack, and additionall takes a parameter and drops that many additional stack items (that's kind of "extra" function - not strictly related to the frame system).
2024-05-27 19:30:53 <KipIngram   > So in between { and } you can use index off of FP in a fixed and immovable way.
2024-05-27 19:31:41 <KipIngram   > I don't use any sort of indentation with it, though - most of my definitions are one-liners.
2024-05-27 19:31:50 <KipIngram   > 45-50 chars.
2024-05-27 19:45:08 <KipIngram   > The main purpose of the thing was to allow me to index into the stack on occasions where I just couldn't figure out a ood way to avoid it.  But the bit where it restores the stack pointer turns out to be a really  nice way to clean up the stack - even if I've got several return paths with different amounts of junk on the stack, it cleans it up properly regardless.
2024-05-27 19:45:26 <KipIngram   > I wasn't really "expecting" that perk, but I've quite enjoyed it.
2024-05-27 21:26:37 <veltas      > crc: Is it possible to get a zip of all the IRC logs?
2024-05-27 21:29:48 <Al2O3       > like every single day of ##forth logs?
2024-05-27 21:29:55 <crc         > forth.chat/allctar
2024-05-27 21:30:42 <crc         > forth.chat/all.tar.gz
2024-05-27 21:30:44 <Al2O3       > http://forth.chat/
2024-05-27 21:31:10 <Al2O3       > compressed,amazingly 48 megs
2024-05-27 21:31:22 <Al2O3       > text is such a nice 'format'
2024-05-27 21:31:25 <crc         > that has all the ##forth and #retro from liberav and the old #forth and #retro from freenode
2024-05-27 21:41:47 <veltas      > Thanks
2024-05-27 21:48:00 <veltas      > Was just looking at old logs with my wife, she was wondering when I joined this channel, she seems to remember it was from near when we got together
2024-05-27 21:48:05 <veltas      > And she's right
2024-05-27 21:48:26 <veltas      > Near start of lockdowns
2024-05-27 21:49:09 <veltas      > Was just remembering also there used to be a good ZX Spectrum IRC channel but there's no good ones now I think, lots of good Z80 assembly advice from them
2024-05-27 21:49:20 <veltas      > I also got good general assembly advice from 6502 people
2024-05-27 21:53:25 <veltas      > About half the regulars seem to have moved on since then
2024-05-27 21:59:58 <Al2O3       > veltas hope your ##forth love life is home plate.
2024-05-27 22:00:09 <Al2O3       > not 1st, 2nd or 3rd base.
2024-05-27 22:00:17 <Al2O3       > :)
2024-05-27 22:01:15 <thrig       > base 10, 16, or 8 are more common
2024-05-27 22:02:19 <zelgomer    > all bases are base 10
2024-05-27 22:03:14 <veltas      > Very true
2024-05-27 22:06:50 <veltas      > Started on my Z80 Forth with some help from here
2024-05-27 22:07:21 <veltas      > Less help than most get on here as I actually read Starting Forth :P
2024-05-27 22:28:06 <skvery      > veltas I also exposed myself to Forth on the ZX Spectrum but never understood the power of it back then.  Today I understand Forth to be the best language to expose the bare metal of on-chip peripherals!  Very close to the bare metal without the stupid compile link download cycle...
2024-05-27 22:30:51 <veltas      > I tried a Forth on the speccy and decided I wanted a newer one
2024-05-27 22:31:25 <veltas      > Although I'm sure the one I tried was better
2024-05-27 22:31:39 <veltas      > New for the sake of new
2024-05-27 22:32:44 <Al2O3       > veltas nice.
2024-05-27 22:33:22 <Al2O3       > link loaders are an interesting problem.
2024-05-27 22:33:36 <Al2O3       > prefer forth, or a runtime that is live.
2024-05-27 22:34:06 <Al2O3       > the whole 'shut door and key' for me is an unnecessary delineation.
2024-05-27 23:45:57 <zelgomer    > bad time to start taking interest in z80
2024-05-27 23:51:00 <zelgomer    > i say as i'm just in the beginning stages of targetting a mips powered pic32
2024-05-27 23:53:22 <MrMobius    > you can always emulate a z80 on the pic32
2024-05-27 23:53:30 <MrMobius    > it's probably faster than the original