2024-07-22 02:02:26 <traveler    > lf94: I appreciate anything that goes beyond the 1 2 + . demo of forth. The Sam Falvo posted by Andreas Wagner on YT is a stand-out in that regard.
2024-07-22 02:03:06 <traveler    > I think microcontroller applications are the most interesting use-case, though.
2024-07-22 03:32:05 <KipIngram   > Love Forth on embedded systems.  It's like what it was born for.
2024-07-22 10:58:27 <veltas      > I agree with traveler, anything is better than nothing, there's not a ton of content
2024-07-22 10:58:51 <veltas      > Streaming might be good too, there are lots of coding streamers but no Forth streamers I don't think
2024-07-22 10:59:13 <veltas      > Literally anything though, please submit in here and I will check it out
2024-07-22 14:59:22 <zelgomer    > who tf wants to sit there and watch somebody else program in real time? i never understood that.
2024-07-22 14:59:31 <zelgomer    > gen z, man
2024-07-22 15:01:45 <veltas      > Gen Z didn't start that trend
2024-07-22 15:02:02 <veltas      > It's the sort of thing you half-watch for a little and then maybe have in the background for company
2024-07-22 15:59:22 <dlowe       > I should start making coding reaction videos
2024-07-22 16:00:18 <dlowe       > char *foo = malloc(  /cut/ I know what you're thinking... are they going to remember to free it?  Let's find out /cut/
2024-07-22 16:00:58 <dlowe       > free(foo); *type type* important_func(*foo)  /cut/ OMGGGGGGGGGGGGGGGGGG
2024-07-22 16:01:19 <dlowe       > tiktok sensation, guaranteed
2024-07-22 18:18:26 <pkal        > can someone give me an example of how to use >number?  (I am trying to read a number from the input, and am considering that or to write my own word instead)
2024-07-22 18:24:08 <KipIngram   > If you do NUMBER from scratch, it gets fairly involved.  Particularly if you are doing something with the decimal point.  Basically you have to notice and "remember" the possible - sign at the beginning, do an unsigned conversion of the number, and then negate it if that - was present.  I do it as a state machine, but mine's extra fancy and will permit base specification via <base>:<digits>
2024-07-22 18:25:08 <GeDaMo      > 0 0 s" 123" >number
2024-07-22 18:25:18 <KipIngram   > However you handle alternate bases, once you're beyond base 10 conversion of ascii digits to digit "values" requires some minor hoop jumping - you have to know to skip over the characters between 9 and A.
2024-07-22 18:25:49 <KipIngram   > To convert 0-9 to a value, just subtract 48.   But for A and above, you have to subtract a bit more.
2024-07-22 18:27:06 <Zarutian_iPad> opinion on base markers like 0x 0b 0o and 0q ?
2024-07-22 18:28:06 <zelgomer    > i use a generic radix prefix notation. 8#123 for octal, for example
2024-07-22 18:28:58 <zelgomer    > and just to be difficult, my default base is hex. so if you want to write a number in decimal, you have to write a#123
2024-07-22 18:29:55 <zelgomer    > i also allow commas in numbers so it's a little nicer on the eyes to write bitmasks in binary:  2#0000,1000,0011,0101
2024-07-22 18:30:42 <Zarutian_iPad> I use _ underscores as verilog and vhdl does
2024-07-22 18:31:49 <zelgomer    > those languages have to use an underscore because the comma is a token. i figured, forth isn't limited by that, so why not embrace it?
2024-07-22 18:31:58 <pkal        > KipIngram: What I had in mind was just to read in natural numbers, so the main problem would be overflowing (then again, I am just playing around so it isn't that important)
2024-07-22 18:32:30 <Zarutian_iPad> because in Iceland and other nordic a comma is the decimal 'point'
2024-07-22 18:32:48 <pkal        > GeDaMo: Thanks, that works with strings, now I just have to combine it with accept.
2024-07-22 18:32:59 <KipIngram   > Ah.  That spares you putzing with the - sign, then.
2024-07-22 18:33:03 <KipIngram   > And any other frills.
2024-07-22 18:33:07 <darkling    > Commas are quite Anglo-centric. Many other countries switch , and . (so . is the separator and , is the decimal point).
2024-07-22 18:33:09 <zelgomer    > there are no countries other than the US
2024-07-22 18:33:12 <Zarutian_iPad> comma in numbers then just looks wierd to me
2024-07-22 18:33:54 <Zarutian_iPad> zelgomer: United Soviets of America?
2024-07-22 18:33:56 <KipIngram   > If you're going strictly base 10 it should be a pretty easy affair.  For overflow I'd just check before adding or multiplying by 10 that my result wouldn't overflow.
2024-07-22 18:34:02 <darkling    > (And in some places, the comma is a scale separator, but not in thousands).
2024-07-22 18:34:07 <zelgomer    > so when you write c, does it bother you that you have to use , to separate arguments and . to write floats?
2024-07-22 18:35:11 <KipIngram   > I.e., when it's time to multiply by 10 make sure your number so far is less than 6554.
2024-07-22 18:35:26 <KipIngram   > Personally I don't worry about overflow.
2024-07-22 18:35:38 <KipIngram   > I expect myself to know what I'm doing when I use it later.
2024-07-22 18:39:50 <pkal        > KipIngram: Is just trying to multiply a safe way to check an overflow in forth?
2024-07-22 19:16:04 <KipIngram   > I've never used a system that checks for overflow on arithmetic operations.
2024-07-22 19:16:35 <KipIngram   > The ones I've used basically just do mod 2^N math and leave it at that.
2024-07-22 19:20:46 <dlowe       > common lisp and scheme do.  They transparently upgrade to bignums
2024-07-22 19:21:04 <dlowe       > so it's never incorrect, your program just gets mysteriously slower
2024-07-22 19:24:30 <darkling    > So does Python.
2024-07-22 19:33:17 <dlowe       > I though Python is always bignums
2024-07-22 19:35:04 <rendar      > my forth can do math operations by wrapping (like C), by saturating, and by throwing error on overflow
2024-07-22 19:35:31 <rendar      > e.g.  1 1 + <-wrappng     1 1 +s <-saturating   1 1 +o  <-overflow
2024-07-22 19:35:47 <dlowe       > makes sense
2024-07-22 19:35:52 <rendar      > :)
2024-07-22 19:44:08 <pkal        > I ended up with this definition https://0x0.st/XpjY.f, but was wondering if there was some way I could allocate the memory on the stack (something like auto char buffer[80]; in C) and discard it?
2024-07-22 19:46:54 <GeDaMo      > Could you use the pad area for that?
2024-07-22 19:49:52 <pkal        > pad area?
2024-07-22 19:50:49 <GeDaMo      > https://forth-standard.org/standard/core/PAD
2024-07-22 19:51:30 <GeDaMo      > "The size of the scratch area whose address is returned by PAD shall be at least 84 characters." https://forth-standard.org/standard/usage#usage:transient
2024-07-22 20:00:58 <pkal        > ah, very useful!  thanks!
2024-07-22 20:12:06 <pkal        > now I have https://0x0.st/Xpjh.f, but I wonder if there is a better way to express the last line?  thinking in lisp, I'd maybe write a macro that takes a list of symbols and numbers indicating how to re-order the stack.
2024-07-22 20:13:01 <GeDaMo      > There's a 2drop
2024-07-22 20:15:12 <crc         > perhaps `2drop drop nip`
2024-07-22 20:15:19 <GeDaMo      > Are you just dropping the value left by accept? You could drop that there
2024-07-22 20:16:36 <GeDaMo      > You could use D>S to go from a double cell value ot a single
2024-07-22 20:24:07 <pkal        > ah, I forgot about the other value by accept, dropping that right away is certainly cleaner
2024-07-22 20:24:34 <pkal        > or actually I can pass that directly to >number