2021-11-25 10:04:43 <crc         > the log bot crashed during the netsplit; missing messages are at http://forth.works/share/1da23941437226823846417de8e84dd5
2021-11-25 14:02:49 <neuro_sys_  > dave0: Is your Forth token threaded?
2021-11-25 15:26:58 <logand`     > in gforth, require ansi.fs gives me: require ansi.fs redefined pn  redefined ;pn  redefined ESC[  redefined at-xy  redefined page   ok
2021-11-25 15:27:21 <logand`     > is that expected?
2021-11-25 16:28:16 <veltas      > logand`: It's not unexpected I don't think
2021-11-25 16:28:33 <dave0       > maw
2021-11-25 16:29:32 <veltas      > dave0: Cool
2021-11-25 16:29:56 <veltas      > Re compiling first FORTH word
2021-11-25 16:30:16 <dave0       > thanks veltas :-)
2021-11-25 16:31:17 <dave0       > there is no compile state yet, i crufted a word that pushed a number on the stack :-)
2021-11-25 16:31:36 <veltas      > Don't go crazy with loops that print numbers from 1 to 10 or anything too fancy
2021-11-25 16:31:36 <dave0       > like    : five 5 literal, ;
2021-11-25 16:31:41 <dave0       > ehehe
2021-11-25 16:31:42 <veltas      > This is FORTH, we're not wizards
2021-11-25 16:32:12 <logand`     > veltas: ok, thanks
2021-11-25 19:09:16 <crc         > dave0: what is the threading model you are using?
2021-11-25 19:09:32 <f-a         > the «thread carefully» model
2021-11-25 19:11:36 <dave0       > crc: indirect threading... it's a 32 bit forth for 8086 realmode... eventually it will be a bootloader for bios
2021-11-25 19:12:29 <crc         > cool. I haven't explored indirect threading so far
2021-11-25 19:12:31 <dave0       > i am testing it with dosbox, and it runs as a .com file in msdos
2021-11-25 19:13:12 <dave0       > crc: it's interesting :-)
2021-11-25 19:14:38 <dave0       > crc: the least standard part is that words are made up of segment:offset pointers, so compiling xt's means running a segment> word that turns a flat address into segment:offset
2021-11-25 19:15:54 <dave0       > but the nice thing is that (apart from compiled pointers) it's a flat memory model... there's 512K of ram and you can make up to that size data structures
2021-11-25 19:16:04 <dave0       > no 64k limits
2021-11-25 19:16:50 <crc         > nice
2021-11-25 19:18:16 <dave0       > i am rewriting INTERPRET so it can compile words.. it's a mass of IF/ELSE/THEN so it's tricky to compile by hand :-p
2021-11-25 19:27:21 <dave0       > : compile,   >segment , ;     >segment is a primitive that takes a 20 bit address and produces segment:offset .. since cells are 32 bits, the values fit nicely
2021-11-25 19:31:48 <dave0       > the dictionary is a linked list, but the link field is a 20 bit address... so i can in theory use the top 12 bits for flags... so far i just use one bit for IMMEDIATE
2021-11-25 19:34:14 <dave0       > also it requires alignment to 4 bytes for some edge cases in ! and @
2021-11-25 19:34:49 <dave0       > so maybe people used to msdos forths might find it tricky
2021-11-25 19:35:13 <dave0       > cos 8086 is wel known for not needing alignment
2021-11-25 19:57:22 <dave0       > oh my god
2021-11-25 19:57:44 <dave0       > : square dup * ;   5 square .    works!
2021-11-25 19:57:52 <dave0       > 25 !
2021-11-25 19:58:58 <dave0       > 
ACTION happy dance

2021-11-25 19:59:03 <dave0       > it's working!
2021-11-25 19:59:09 <dave0       > now strings so i can hello world
2021-11-25 20:00:04 <f-a         > very good
2021-11-25 20:03:43 <dave0       > thanks f-a :-)
2021-11-25 21:24:40 <dave0       > nice
2021-11-25 21:24:44 <dave0       > hello world
2021-11-25 21:24:46 <dave0       > no worries
2021-11-25 21:32:36 <dave0       > ok time for lunch bbl