X-Git-Url: https://git.rrq.au/?a=blobdiff_plain;f=main.asm;h=3da2e871d8d4ccc7767f7356ebb057c815b9201b;hb=b1e9665f4639d9fe9ef0a3af7cc09c68c49b928b;hp=e5444b87a58ba20ad0f984627d05ef0e99efecec;hpb=ad2d7300c0ec7b547259fdfd682bb7f0ee6a01ab;p=rrq%2Fjonasforth.git

diff --git a/main.asm b/main.asm
index e5444b8..3da2e87 100644
--- a/main.asm
+++ b/main.asm
@@ -44,6 +44,10 @@ docol:
 
 ;; This word is called at the end of a Forth definition. It just needs to
 ;; restore the old value of RSI (saved by 'docol') and resume execution.
+EXIT_entry:
+  dq 0
+  db 4
+  db 'EXIT'
 EXIT:
   dq .start
 .start:
@@ -52,6 +56,10 @@ EXIT:
 
 ;; LIT is a special word that reads the next "word pointer" and causes it to be
 ;; placed on the stack rather than executed.
+LIT_entry:
+  dq EXIT_entry
+  db 3
+  db 'LIT'
 LIT:
   dq .start
 .start:
@@ -59,6 +67,78 @@ LIT:
   push rax
   next
 
+;; Given a string (a pointer following by a size), return the location of the
+;; dictionary entry for that word. If no such word exists, return 0.
+FIND_entry:
+  dq LIT_entry
+  db 4
+  db 'FIND'
+FIND:
+  dq .start
+.start:
+  mov [.rsi], rsi
+  pop [.search_length]
+  pop [.search_buffer]
+
+  ;; RSI contains the entry we are currently looking at
+  mov rsi, [latest_entry]       ; Start with the last added word
+
+.loop:
+  movzx rcx, byte [rsi + 8]     ; Length of word being looked at
+  cmp rcx, [.search_length]
+  jne .next    ; If the words don't have the same length, we have the wrong word
+
+  ;; Otherwise, we need to compare strings
+  lea rdx, [rsi + 8 + 1]        ; Location of character being compared in entry
+  mov rdi, [.search_buffer]     ; Location of character being compared in search buffer
+.compare_char:
+  mov al, [rdx]
+  mov ah, [rdi]
+  cmp al, ah
+  jne .next                     ; They don't match; try again
+  inc rdx                       ; These characters match; look at the next ones
+  inc rdi
+  loop .compare_char
+
+  jmp .found                    ; They match! We are done.
+
+.next:
+  mov rsi, [rsi]                ; Look at the previous entry
+  cmp rsi, 0
+  jnz .loop                    ; If there is no previous word, exit and return 0
+
+.found:
+  push rsi
+
+  mov rsi, [.rsi]
+  next
+
+;; BRANCH is the fundamental mechanism for branching. BRANCH reads the next word
+;; as a signed integer literal and jumps by that offset.
+BRANCH_entry:
+  dq FIND_entry
+  db 6
+  db 'BRANCH'
+BRANCH:
+  dq .start
+.start:
+  add rsi, [rsi] ; [RSI], which is the next word, contains the offset; we add this to the instruction pointer.
+  next           ; Then, we can just continue execution as normal
+
+;; 0BRANCH is like BRANCH, but it jumps only if the top of the stack is zero.
+ZBRANCH:
+  dq .start
+.start:
+  ;; Compare top of stack to see if we should branch
+  pop rax
+  cmp rax, 0
+  jnz .dont_branch
+.do_branch:
+  jmp BRANCH.start
+.dont_branch:
+  add rsi, 8     ; We need to skip over the next word, which contains the offset.
+  next
+
 ;; Expects a character on the stack and prints it to standard output.
 EMIT:
   dq .start
@@ -82,9 +162,20 @@ NEWLINE:
   dq EMIT
   dq EXIT
 
+;; Prints a space to standard output.
+SPACE_entry:
+  dq BRANCH_entry
+  db 5
+  db 'SPACE'
+SPACE:
+  dq docol
+  dq LIT, ' '
+  dq EMIT
+  dq EXIT
+
 ;; Read a word from standard input and push it onto the stack as a pointer and a
 ;; size. The pointer is valid until the next call to READ_WORD.
-READ_WORD:  ; 400170
+READ_WORD:
   dq .start
 .start:
   mov [.rsi], rsi
@@ -136,9 +227,63 @@ READ_WORD:  ; 400170
 
   next
 
+;; Takes a string on the stack and replaces it with the decimal number that the
+;; string represents.
+PARSE_NUMBER:
+  dq .start
+.start:
+  pop [.length]                 ; Length
+  pop rdi                       ; String pointer
+  mov r8, 0                     ; Result
+
+  ;; Add (10^(rcx-1) * parse_char(rdi[length - rcx])) to the accumulated value
+  ;; for each rcx.
+  mov rcx, [.length]
+.loop:
+  ;; First, calcuate 10^(rcx - 1)
+  mov rax, 1
+
+  mov r9, rcx
+  .exp_loop:
+    dec r9
+    jz .break
+    mov rbx, 10
+    mul rbx
+    jmp .exp_loop
+  .break:
+
+  ;; Now, rax = 10^(rcx - 1).
+
+  ;; We need to calulate the value of the character at rdi[length - rcx].
+  mov rbx, rdi
+  add rbx, [.length]
+  sub rbx, rcx
+  movzx rbx, byte [rbx]
+  sub rbx, '0'
+
+  ;; Multiply this value by rax to get (10^(rcx-1) * parse_char(rdi[length - rcx])),
+  ;; then add this to the result.
+  mul rbx
+
+  ;; Add that value to r8
+  add r8, rax
+
+  dec rcx
+  jnz .loop
+
+  push r8
+
+  next
+
+READ_NUMBER:
+  dq docol
+  dq READ_WORD
+  dq PARSE_NUMBER
+  dq EXIT
+
 ;; Takes a string (in the form of a pointer and a length on the stack) and
 ;; prints it to standard output.
-TYPE:
+TELL:
   dq .start
 .start:
   mov rbx, rsi
@@ -179,6 +324,10 @@ PUSH_YOU_TYPED:
   push you_typed_string.length
   next
 
+HELLO_entry:
+  dq SPACE_entry
+  db 5
+  db 'HELLO'
 HELLO:
   dq docol
   dq LIT, 'H', EMIT
@@ -190,23 +339,110 @@ HELLO:
   dq NEWLINE
   dq EXIT
 
+;; .U prints the value on the stack as an unsigned integer in hexadecimal.
+DOTU_entry:
+  dq HELLO_entry
+  db 2
+  db '.U'
+DOTU:
+  dq .start
+.start:
+  mov [.length], 0
+  mov [.printed_length], 1
+  pop rax                       ; RAX = value to print
+  push rsi                      ; Save value of RSI
+
+  ;; We start by constructing the buffer to print in reverse
+
+.loop:
+  mov rdx, 0
+  mov rbx, $10
+  div rbx                       ; Put remainer in RDX and quotient in RAX
+
+  ;; Place the appropriate character in the buffer
+  mov rsi, .chars
+  add rsi, rdx
+  mov bl, [rsi]
+  mov rdi, .rbuffer
+  add rdi, [.length]
+  mov [rdi], bl
+  inc [.length]
+
+  ;; .printed_length is the number of characters that we ulitmately want to
+  ;; print. If we have printed a non-zero character, then we should update
+  ;; .printed_length.
+  cmp bl, '0'
+  je .skip_updating_real_length
+  mov rbx, [.length]
+  mov [.printed_length], rbx
+.skip_updating_real_length:
+
+  cmp [.length], 16
+  jle .loop
+
+  ;; Flip buffer around, since it is currently reversed
+  mov rcx, [.printed_length]
+.flip:
+  mov rsi, .rbuffer
+  add rsi, rcx
+  dec rsi
+  mov al, [rsi]
+
+  mov rdi, .buffer
+  add rdi, [.printed_length]
+  sub rdi, rcx
+  mov [rdi], al
+
+  loop .flip
+
+  ;; Print the buffer
+  mov rax, 1
+  mov rdi, 1
+  mov rsi, .buffer
+  mov rdx, [.printed_length]
+  syscall
+
+  ;; Restore RSI and continue execution
+  pop rsi
+  next
+
 MAIN:
   dq docol
   dq HELLO
-  dq READ_WORD
-  dq LIT, you_typed_string
-  dq LIT, you_typed_string.length
-  dq TYPE
-  dq TYPE
-  dq NEWLINE
-  dq HELLO
+  dq LIT, SPACE_entry, DOTU, NEWLINE
+  dq LIT, HELLO_entry, DOTU, NEWLINE
+  dq LIT, DOTU_entry, DOTU, NEWLINE
+  dq LIT, SPACE_string, LIT, SPACE_string.length, TELL, SPACE
+  dq LIT, SPACE_string, LIT, SPACE_string.length, FIND, DOTU, NEWLINE
+  dq LIT, HELLO_string, LIT, HELLO_string.length, TELL, SPACE
+  dq LIT, HELLO_string, LIT, HELLO_string.length, FIND, DOTU, NEWLINE
+  dq LIT, DOTU_string, LIT, DOTU_string.length, TELL, SPACE
+  dq LIT, DOTU_string, LIT, DOTU_string.length, FIND, DOTU, NEWLINE
+  dq LIT, HELLA_string, LIT, HELLA_string.length, TELL, SPACE
+  dq LIT, HELLA_string, LIT, HELLA_string.length, FIND, DOTU, NEWLINE
   dq TERMINATE
 
 segment readable writable
 
+latest_entry dq DOTU_entry
+
+SPACE_string db 'SPACE'
+.length = $ - SPACE_string
+HELLO_string db 'HELLO'
+.length = $ - HELLO_string
+DOTU_string db '.U'
+.length = $ - DOTU_string
+HELLA_string db 'HELLA'
+.length = $ - HELLA_string
+
+
 you_typed_string db 'You typed: '
 .length = $ - you_typed_string
 
+FIND.search_length dq ?
+FIND.search_buffer dq ?
+FIND.rsi dq ?
+
 READ_WORD.rsi dq ?
 READ_WORD.rax dq ?
 READ_WORD.max_size = $FF
@@ -214,6 +450,14 @@ READ_WORD.buffer rb READ_WORD.max_size
 READ_WORD.length db ?
 READ_WORD.char_buffer db ?
 
+DOTU.chars db '0123456789ABCDEF'
+DOTU.buffer rq 16               ; 64-bit number has no more than 16 digits in hex
+DOTU.rbuffer rq 16
+DOTU.length dq ?
+DOTU.printed_length dq ?
+
+PARSE_NUMBER.length dq ?
+
 ;; Return stack
 rq $2000
 return_stack_top: