Commit Graph

17 Commits

Author SHA1 Message Date
Andrius Štikonas 279941646f
Fix a typo:
Recieved -> Received
2019-10-23 20:01:23 +01:00
Fotis Georgatos e95dc40848
typo: recieved -> received 2019-06-25 00:16:48 +02:00
Jeremiah Orians ed2dc916dd
Minor update for get_machine.c 2018-11-07 21:53:48 -05:00
Jeremiah Orians ea9748946e
Upgraded M2-Planet and Changelog to reflect improvements 2018-10-17 18:25:11 -04:00
Jeremiah Orians 67cc824a40
Update to current M2-Planet and add janneke's clone 2018-08-31 19:08:06 -04:00
Jeremiah Orians 5c9aa4ab15
Fix license header typo 2018-08-26 21:50:34 -04:00
Jeremiah Orians 2988d0e391
Release_0.2.0 2018-08-26 12:12:30 -04:00
Jeremiah Orians c06faa77e0
Release 0.0.9 2017-08-18 12:02:02 -04:00
Jeremiah Orians 4c187f68b4
Ensured compliance with new stage0 assembler standard 2017-08-04 21:05:33 -04:00
Jeremiah Orians efa8689570
Update changelog and add missing copyright notice 2017-07-13 21:16:41 -04:00
Jeremiah Orians 228c8b20e2
Function to help beginners 2017-07-13 20:51:57 -04:00
Jeremiah Orians e113458a62
Made JUMP-TO use 32bit absolute addresses 2017-07-13 18:24:26 -04:00
Jeremiah Orians 69a77610cf
Minor Forth Tweak 2017-07-12 21:07:53 -04:00
Caleb Ristvedt 1a6c7d5afe
Fix CREATE, implement DOES>, more words
Implemented some new words in forth.s:
* 'DODOES
  Gives the address of the assembly for DODOES. Used by DOES> when compiling
  machine code to jump to it and activate the following high-level forth
  code. This is how non-primitive actions for a word can be set.
* LATEST!
  Sets the LATEST pointer. For when the user wants to write their own way of
  extending the dictionary (word lists, vocabularies, etc).
* R@
  Copies the top of the return stack to the data stack. Useful for when loops
  start getting non-trivial and you start wondering of all those ROTs aren't
  efficient enough. Maybe someday we'll implement a register-allocating
  compiler. Probably not. Too complicated for bootstrapping.
* Not really words but used: DODOES and DOVAR.
  DOVAR is the default behavior of CREATEd words - it just pushes a pointer to
  whatever was allotted right after the word was created. DODOES is jumped to in
  order to invoke high-level forth behaviors for words.
* Changed behavior of CREATE to better match the way it usually works.

Added some words to inital_library.fs (we should fix that name sometime):
* Added a variable BASE to control what base numbers are printed in.
* Turns out I misunderstood what WORD did and mis-diagnosed an early problem in
  ' - so it turns out that nothing was being allotted in the dictionary, just
  stuff in "string space", which if I understand properly just sort of fills up
  infinitely as long as WORD is being invoked. Some other forths try to solve
  this by introducing an input buffer for the current line, at the cost of some
  complexity (as words like [CHAR] and S" have to take that into
  account). Anyway, the HERE and DP! are gone from ' now.
* Added BOUNDS for common setup for sequence-iterating loops.
* Added region-comment "(" (note that it doesn't nest), used most often for
  stack comments.
* Added hex dump printer DUMP and support words, number of bytes printed per
  line is controlled by LINE-SIZE. It looks pretty nice, much of the design is
  based off of gforth's.
* Added DOES> and supporting words, and used it to make VARIABLE, CONSTANT, and
  DEFER.
* Added TUCK, MIN, SPACES, :NONAME, FILL, and <>.

Add LSHIFT and RSHIFT to the wishlist, as ghetto shifting with division seems to
behave a bit strangely when given negative values.

Also, we're now advanced enough to support the tictactoe I wrote awhile back for
gforth! http://paste.lisp.org/display/349394
2017-06-24 07:40:22 -04:00
Caleb Ristvedt c159b6a927
Refactor, fix bugs, and reword comments in initial_library.fs
Some of the comments seemed a bit off. Additionally, there was a bug in AGAIN
in which it didn't actually compile a jump. Also, WHILE's compile-time stack
effect is now consistent with gforth, which means we can have multi-exit while
loops. Also, added UNTIL. Also, added [CHAR] so we don't have to keep looking up
ascii values for characters. Also, whichever bug it was that caused problems
with splitting definitions up across multiple lines is now gone, so long live
the newlines!
2017-06-21 05:47:43 -04:00
Jeremiah Orians b3eedb60de
Refining FORTH to enable more productive development 2017-06-17 00:02:33 -04:00
Jeremiah Orians 1e28eadc7a
Preparing the our next release with our first stage3 program 2017-06-16 23:02:58 -04:00