Migrate nlTokBuilder + nlTokStream -> nlLexer

src/nlx.nim

@@ -1,19 +1,19 @@
 import os
 import noether/lib/io
-import noether/lexer/[tok, tokstream]
+import noether/lexer/tokbuilder
 # import noether/parser/parser
 
 {.hint: "Don't forget to drink more water (^_^)".}
 when isMainModule:
   echo "Noether Lang Extras v0.1.0 - nlx"
 
-  var inStream = if paramCount() > 0: streamFile(paramStr 1)
-                 else: streamString(readAll stdin)
+  var stream = if paramCount() > 0: streamFile(paramStr 1)
+               else: streamString(readAll stdin)
 
-  var stream = newTokStream(inStream)
+  var lexer = newLexer(stream)
   # # DumpTok
-  while stream.progress():
-    echo stream.tok
+  while lexer.progress():
+    echo lexer.tok
 
   # DumpTree
   # discard parse(tokStream)

src/noether/lexer/tokbuilder.nim

@@ -8,9 +8,11 @@ export tok
 type
   # Abstracts the "building process" (lexing) 
   # of nlTok objects from a given Stream of characters.
-  nlTokBuilder* = object
+  nlLexer* = object
     stream: Stream
-    tok: nlTok # the build token
+    done*: bool
+    tok*: nlTok # new finished token
+    btok: nlTok # the build token
     # track line number, line content, etc
     line: string
     lineNum: int
@@ -19,18 +21,18 @@ type
     char: char
     cTKind: nlTokKind
 
-proc atEOL(builder: nlTokBuilder): bool {.inline.} =
-  result = (builder.char == '\n')
-proc atEOF(builder: nlTokBuilder): bool {.inline.} =
-  result = (builder.char == '\0')
+proc atEOL(lexer: nlLexer): bool {.inline.} =
+  result = (lexer.char == '\n')
+proc atEOF(lexer: nlLexer): bool {.inline.} =
+  result = (lexer.char == '\0')
 
-# Initialise a new token builder
-proc newBuilder(stream: var Stream): nlTokBuilder =
-  # NOTE: initial builder.char value is arbitrary, 
-  # NOTE: but CANNOT be initialised to the default '\0'
-  result = nlTokBuilder(
+# Initialise a new lexer
+proc newLexer*(stream: var Stream): nlLexer =
+  result = nlLexer(
     stream: stream,
+    done: false,
     tok: emptyTok(0),
+    btok: emptyTok(0),
     line: "",
     lineNum: 1,
     pos: -1,    # after initial readChar this -> 0
@@ -39,70 +41,104 @@ proc newBuilder(stream: var Stream): nlTokBuilder =
 
 
 #[ ====================================================== ]
- | nlTokBuilder Internal Interface for Token Construction ]
+ | nlLexer Internal Interface for Token Construction ]
  ]#
   
 # Reset the build token to be "empty"
-proc resetBuild(builder: var nlTokBuilder) =
-  builder.tok = emptyTok(builder.pos)
+proc resetBuild(lexer: var nlLexer) =
+  lexer.btok = emptyTok(lexer.pos)
 
 # "Finishes" the build token by setting various properties
-proc finishBuild(builder: var nlTokBuilder) =
-  builder.tok.lineNum = builder.lineNum
-  builder.tok.endPos = builder.pos
-  builder.tok.lit = builder.line[builder.tok.startPos ..< builder.line.high]
+proc finishBuild(lexer: var nlLexer) =
+  lexer.btok.lineNum = lexer.lineNum
+  lexer.btok.endPos = lexer.pos
+  lexer.btok.lit = lexer.line[lexer.btok.startPos ..< lexer.line.high]
 
 # Finish, return, and reset the build token
-proc flushBuild(builder: var nlTokBuilder): nlTok = 
-  finishBuild(builder)
-  result = builder.tok
-  resetBuild(builder)
+proc flushBuild(lexer: var nlLexer): nlTok = 
+  finishBuild(lexer)
+  result = lexer.btok
+  resetBuild(lexer)
 
 # Is the build token "compatible" with the current char? (if not then flushbuild)
 # NOTE: This implicitly handles Windows CRLF, Unix LF, & Mac OS CR compatability 
 # NOTE: since atEOL => '\n', but '\r' and '\n' are both tkEOL so they both flush.
-proc isIncompatibleBuild(builder: nlTokBuilder): bool =
-  result = (builder.cTKind != builder.tok.kind or builder.atEOL())
+proc isIncompatibleBuild(lexer: nlLexer): bool =
+  result = (lexer.cTKind != lexer.btok.kind or lexer.atEOL())
 
 # Inherit the build token's type from current char
-proc inherit(builder: var nlTokBuilder) = 
-  builder.tok.kind = builder.cTKind
+proc inherit(lexer: var nlLexer) = 
+  lexer.btok.kind = lexer.cTKind
 
-# Add a character to the nlTokBuilder's build token.
+# Add a character to the nlLexer's build token.
 # Flushes and returns the build token if finished.
-proc appendBuild(builder: var nlTokBuilder): Option[nlTok] =    
+proc appendBuild(lexer: var nlLexer): Option[nlTok] =    
   # untyped build tokens inherit type immediately
-  if builder.tok.isUntyped():
-    builder.inherit()
+  if lexer.btok.isUntyped():
+    lexer.inherit()
   
   # check character and build token compatability
-  if isIncompatibleBuild(builder):
+  if isIncompatibleBuild(lexer):
       # flush old build token, the new one inherits type
-      result = some(flushBuild(builder))
-      builder.inherit()
+      result = some(flushBuild(lexer))
+      lexer.inherit()
   else:
     result = none(nlTok)
 
-#[ ========================================== ]
- | nlTokBuilder Char Stream Reading Interface ]
+#[ ========================================= ]
+ | nlLexer Internal Char Streaming Interface ]
  ]#
       
 # Read the next char in the stream
 # NOTE: readChar raises IOError on error, returns \0 on EOF
-proc readChar*(builder: var nlTokBuilder): bool =
-  if builder.atEOL():
-    inc builder.lineNum
-  # sets builder.char to '\0' if EOF
-  builder.char = builder.stream.readChar()
-  builder.cTKind = getTokKind(builder.char)
-  builder.line.add(builder.char)
-  inc builder.pos
-  result = builder.atEOF()
+proc readChar(lexer: var nlLexer): bool =
+  if lexer.atEOL():
+    inc lexer.lineNum
+  # sets lexer.char to '\0' if EOF
+  lexer.char = lexer.stream.readChar()
+  lexer.cTKind = getTokKind(lexer.char)
+  lexer.line.add(lexer.char)
+  inc lexer.pos
+  result = lexer.atEOF()
 
+#[ ========================
+ | nlLexer Public Interface
+ ]#
+  
 # Read until EOL and return the current line
-# NOTE: Does NOT update the builder's state (unsafe)
+# NOTE: Does NOT update the lexer's state (unsafe)
 # NOTE: ONLY call if a lex/parse error needs displaying
-proc unsafeGetLine*(builder: var nlTokBuilder): string =
-  while not builder.atEOL() and builder.readChar():
+proc unsafeGetLine*(lexer: var nlLexer): string =
+  while not lexer.atEOL() and lexer.readChar():
     discard
-  result = builder.line
+  result = lexer.line
+
+# Lexes and returns the next token in the "token stream"
+# via repeatedly calling readChar() and appendBuild().
+# Returns a boolean indicating whether EOF has been reached.
+# NOTE: access the new token via `stream.tok`
+proc progress*(lexer: var nlLexer): bool =
+  # Return prematurely if already closed
+  if lexer.done:
+    return false
+  while true:
+    let 
+      atEOF = lexer.readChar()
+      flushedTok = lexer.appendBuild()
+      newTokBuilt = flushedTok.isSome
+    
+    if newTokBuilt:
+      lexer.tok = flushedTok.get()
+    # if canProgress and atEOF:
+    # if atEOF:
+    #   if newTokBuilt:
+    #     stream.isClosed = true
+    #   return newTokBuilt
+    # elif newTokBuilt:
+    #   return true
+    if newTokBuilt:
+      if atEOF:
+        lexer.done = true
+      return true
+    elif atEOF:
+      return false