2024-11-14 09:00:55 <veltas      > lilo_booter: To iterate through floats you can increment them as an integer (or decrement if negative)
2024-11-14 09:01:39 <veltas      > Assuming they're IEEE 754
2024-11-14 09:03:57 <veltas      > Oh I see that's not what you were asking lol
2024-11-14 09:17:57 <veltas      > I would use BEGIN / WHILE / REPEAT :  3.14159265359E FCONSTANT PI  FVARIABLE X   : MY-LOOP   PI FNEGATE X F!  BEGIN  X F@  PI  F< WHILE  ....  REPEAT ;
2024-11-14 09:20:13 <veltas      > Making ?FDO ... FLOOP with the standard word set would be painful because there's no way to put floats on the return stack portably
2024-11-14 09:20:29 <veltas      > At least not in ANS Forth
2024-11-14 09:21:07 <veltas      > A lot of stuff is painful with ANS floats, they're really not a first-class type, they're just there because they know you need it
2024-11-14 09:22:35 <veltas      > You'd think they'd let you use FVARIABLE to start a region of float-aligned addresses, but the standard disclaims this in general
2024-11-14 09:22:48 <veltas      > I would guess most Forths do let you do that though, like gforth
2024-11-14 13:19:29 <xentrac     > hmm, is there any way to start a region of float-aligned addresses in the standard?
2024-11-14 13:19:48 <xentrac     > or to float-align here
2024-11-14 13:20:00 <amby        > float aligned?
2024-11-14 13:20:13 <amby        > oh as in. aligned to the size of a float
2024-11-14 13:20:30 <GeDaMo      > https://forth-standard.org/standard/float/FALIGN
2024-11-14 13:20:43 <xentrac     > aha!
2024-11-14 13:21:38 <xentrac     > not necessarily the size, just the necessary alignment so you don't blow up with a bus error or get unreasonably slow code
2024-11-14 13:21:54 <amby        > yeah
2024-11-14 13:22:30 <xentrac     > I see there's also https://forth-standard.org/standard/float/DFALIGN https://forth-standard.org/standard/float/DFLOATS https://forth-standard.org/standard/float/DFLOATPlus https://forth-standard.org/standard/float/FLOATPlus https://forth-standard.org/standard/float/FLOATS
2024-11-14 13:22:45 <GeDaMo      > Also https://forth-standard.org/standard/float/FALIGNED
2024-11-14 13:22:48 <xentrac     > yes
2024-11-14 13:23:27 <xentrac     > and https://forth-standard.org/standard/float/DFALIGNED
2024-11-14 13:23:49 <xentrac     > so you have everything you need to make structs or arrays with floats or doubles on the heap or in the dictionary
2024-11-14 22:35:04 <Zarutian_iPad> a devious thought occured to me: an js/ts to a forth compiler
2024-11-14 23:22:32 <veltas      > xentrac: But annoyingly there's no FCREATE etc
2024-11-14 23:22:39 <veltas      > So it's a bit lopsided
2024-11-14 23:25:13 <veltas      > I end up doing stuff like  : F, HERE >R  1 FLOATS ALLOT  R> F! ;    FALIGN HERE x F, x F, x F,  CONSTANT MY-ARRAY