2022-10-13 05:26:11 <crcx        > I've been limiting line length to 64 characters (as a side effect of using blocks now), but am tolerant of multiline definitions
2022-10-13 10:53:47 <olle        > \o
2022-10-13 10:54:42 <thrig       > : \o ." HELLO " ;
2022-10-13 11:06:14 <olle        > ."
2022-10-13 11:06:15 <olle        > Hm
2022-10-13 14:25:51 <KipIngram   > Hi olle
2022-10-13 15:27:38 <olle        > Any good links for embedded DSL with Forth?
2022-10-13 15:35:03 <olle        > Maybe everything you do with Forth is DSL?
2022-10-13 15:36:57 <thrig       > Damn Succinct Language
2022-10-13 15:37:06 <olle        > Also a link about type-safety in Forth would be nice
2022-10-13 15:37:15 <olle        > thrig: :)
2022-10-13 15:42:52 <forthmacs   > In Mecrisp Stellaris Forth, all stack items are guaranteed to be 32-bit long binary numbers
2022-10-13 15:42:54 <forthmacs   > :)
2022-10-13 15:45:49 <olle        > baha
2022-10-13 18:23:39 <olle        > gforth is not case sensitive?
2022-10-13 18:23:50 <olle        > Weird
2022-10-13 18:26:25 <thrig       > pretty sure the spec says YOU MUST USE SHOUTY but most all moderns don't
2022-10-13 18:27:13 <olle        > thrig: ok, what was the use case of small letters in the spec, then?
2022-10-13 21:24:15 <remexre     > what i do is upper-case = "standardish or general programming" word, lower-case = domain-specific word
2022-10-13 21:25:05 <thrig       > some only accept lower case (eforth)
2022-10-13 21:37:48 <KipIngram   > type safety in Forth?  Sounds a little off to me...
2022-10-13 21:37:53 <KipIngram   > Forth doesn't have types.
2022-10-13 21:38:05 <thrig       > maybe retroforth has types or something?
2022-10-13 21:38:20 <KipIngram   > Yeah, I guess there are experimental ones.
2022-10-13 21:38:24 <thrig       > (I don't pay typed languages much mind)
2022-10-13 21:38:30 <KipIngram   > I've thought about things, but never done any of them yet.
2022-10-13 21:38:44 <KipIngram   > Seems like a simple system would be fairly straightforward.
2022-10-13 21:39:02 <KipIngram   > And then you could overload word names.
2022-10-13 21:39:17 <thrig       > or have a type prefix
2022-10-13 21:39:26 <KipIngram   > Since the word lookup could match type requirements against the stack contents.
2022-10-13 21:39:43 <KipIngram   > Well, at compile time - that's how I'd thought about doing it.
2022-10-13 21:40:06 <KipIngram   > The compiled code wouldn't deal with any of it.
2022-10-13 21:40:38 <KipIngram   > The main rub I saw was variant code paths having variant type effects on the stack.
2022-10-13 21:40:55 <KipIngram   > But maybe you'd just declare that not allowed.
2022-10-13 21:42:04 <KipIngram   > Compiler would track the type pattern of the stack, and if it compiled a conditional and then the pattern after wasn't the same as the pattern before - throw an error.
2022-10-13 21:47:23 <remexre     > i believe siraben was looking at such a thing
2022-10-13 22:07:45 <crcx        > thrig: the prefixes on names in RetroForth are basically just an alternative to wordlists for grouping related words