From ebef45818626b0ee8a87cd1d7920cbd458295c1a Mon Sep 17 00:00:00 2001
From: Emile Clark-Boman <eclarkboman@gmail.com>
Date: Wed, 18 Jun 2025 19:04:33 +1000
Subject: [PATCH] Implementing greedy matching via template macros (not in a
 working state)

---
 lang/NOTES                       |  2 ++
 src/noether/parser/parser.nim    | 62 ++++++++++++--------------------
 src/noether/parser/parseutil.nim | 58 ++++++++++++++++++++++++++++++
 3 files changed, 83 insertions(+), 39 deletions(-)
 create mode 100644 src/noether/parser/parseutil.nim

diff --git a/lang/NOTES b/lang/NOTES
index 0e8e5c9..df47f66 100644
--- a/lang/NOTES
+++ b/lang/NOTES
@@ -1,3 +1,5 @@
 Todo:
 - [ ] Not currently sure how the lexer will interpret non-latin characters (make sure it handles all unicode)
 - [ ] The lexer currently only handles a limited number of escape codes / whitespace characters
+
+- [ ] Mark most lexer procedures with the {.inline.} pragma (I thought this was active by default)
diff --git a/src/noether/parser/parser.nim b/src/noether/parser/parser.nim
index f83861d..0598075 100644
--- a/src/noether/parser/parser.nim
+++ b/src/noether/parser/parser.nim
@@ -1,49 +1,33 @@
 import strutils
+include parseutil
 
-import nodes
-import ../lexer/tokstream
+# NOTE: Matching between two tokens will fill `node` with everything
+# NOTE: between those two tokens EXCLUDING the two tokens themselves.
+proc parseMatch(tokStream: var nlTokStream, 
+                node: var nlNode,
+                matchType: nlTokType): nlParseStat =
+  result = greed(
+    tokStream, 
+    node.toks, 
+    satisfyMatch(matchType),
+  )
+proc parseMatchLine(tokStream: var nlTokStream, 
+                   node: var nlNode,
+                   matchType: nlTokType): nlParseStat =
+  result = greed(
+    tokStream, 
+    node.toks, 
+    satisfyMatchEOL(matchType),
+  )
 
-type
-  nlParseStat = enum
-    OK,
-    UNMATCHED,
-    TOOBIG,
-
-proc `*`(stat: nlParseStat, b: bool): nlParseStat =
-  result = if b: stat else: nlParseStat.OK
-
-# Greed will consume anything except a punishment
-# Returns a boolean indicating if it succeeded
-proc greed(tokStream: var nlTokStream, toks: var seq[nlTok], satisfy: proc(tok: nlTok): bool): bool =
-  var tok: nlTok
-  while tokStream.nextTok(tok):
-    toks.add(tok)
-    if satisfy(tok):
-      return true
-  result = false
-
-proc greedEOL(tokStream: var nlTokStream, toks: var seq[nlTok], satisfy: nlTokType): bool =
-  var tok: nlTok
-  while tokStream.nextTok(tok):
-    toks.add(tok)
-    if tok.tType == satisfy or tok.tType == nlTokType.EOL:
-      return true
-  result = false
-
-proc satisfyTypeOrEOL(tokType: nlTokType, tok: nlTok): bool = 
-  
-
-proc prsMatchEOL(tokStream: var nlTokStream, toks: var seq[nlTok]): nlParseStat =
-  
-  
-proc parse_strl(tokStream: var nlTokStream, node: var nlNode): nlParseStat =
+proc parseStrL(tokStream: var nlTokStream, node: var nlNode): nlParseStat =
   node = nlNode(
     nType: nlNodeType.STRL
   )
   node.addTok(tokStream.currTok)
   result = nlParseStat.UNCLOSED * not greedEOL(tokStream, node.toks, nlTokType.DQUO)
 
-proc parse_chrl(tokStream: var nlTokStream, node: var nlNode): bool =
+proc parseChrL(tokStream: var nlTokStream, node: var nlNode): bool =
   node = nlNode(
     nType: nlNodeType.CHRL
   )
@@ -59,7 +43,7 @@ proc parse*(tokStream: var nlTokStream): nlNode =
     case tok.tType:
     of nlTokType.DQUO:
       # Attempt to parse string literal
-      if not parse_strl(tokStream, node):
+      if not parseStrL(tokStream, node):
         echo "Unmatched Double Quotation! Malformed String Literal"
         echo tokStream.currLine()
         echo repeat(" ", tok.startPos), '^'
@@ -68,7 +52,7 @@ proc parse*(tokStream: var nlTokStream): nlNode =
         echo node[]
     of nlTokType.SQUO:
       # Attempt to parse string literal
-      if not parse_chrl(tokStream, node):
+      if not parseChrL(tokStream, node):
         echo "Unmatched Single Quotation! Malformed Character Literal"
         echo tokStream.currLine()
         echo repeat(" ", tok.startPos), '^'
diff --git a/src/noether/parser/parseutil.nim b/src/noether/parser/parseutil.nim
new file mode 100644
index 0000000..6fa1243
--- /dev/null
+++ b/src/noether/parser/parseutil.nim
@@ -0,0 +1,58 @@
+import nodes
+import ../lexer/tokstream
+
+type
+  # NOTE: Values above __FAIL__ indicate a failed state
+  nlParseStat* = enum
+    OK,
+    __FAIL__,
+    MIDAS, # Greedy search was never satisfied
+    UNMATCHED,
+    TOOBIG,
+
+proc `*`(stat: nlParseStat, b: bool): nlParseStat =
+  result = if b: stat else: nlParseStat.OK
+
+proc isFail*(stat: nlParseStat): bool = 
+  result = (stat >= nlParseStat.__FAIL__)
+
+
+#[ "Greed" refers to something I mentioned in my discussion on
+ |  Noether's grammar (in an EBNF-like language). Greed just
+ |  means "everything until a condition is satisified".
+ |  That condition should be supplied by a Nim procedural type.
+ ]#
+  
+# Greed will consume anything until a condition is satisfied
+# Returns false if the greed was never satisfied (OMG!!)
+proc greed(tokStream: var nlTokStream, 
+           toks: var seq[nlTok], 
+           satisfy: proc(tok: nlTok): bool,
+           ): nlParseStat =
+  var tok: nlTok
+  while tokStream.nextTok(tok):
+    toks.add(tok)
+    if satisfy(tok):
+      return nlParseStat.OK
+  result = nlParseStat.UNMATCHED
+
+proc greedLine(tokStream: var nlTokStream, 
+               toks: var seq[nlTok], 
+               satisfy: proc(tok: nlTok): bool): nlParseStat =
+  var tok: nlTok
+  while tokStream.nextTok(tok):
+    toks.add(tok)
+    if satisfy(tok):
+      return true
+  result = 
+
+#[ Templates for generating greed satisfying conditions.
+ ]#
+
+# Satisfied if it finds nlTok of type matchType
+template satisfyMatch(matchType: nlTokType) = 
+  proc(tok: nlTok): bool {.inline.} = (tok.tType == matchType)
+  
+# Satisfied if it finds nlTok of type matchType or EOL reached
+template satisfyMatchEOL(matchType: nlTokType) = 
+  proc(tok: nlTok): bool {.inline.} = (tok.tType == matchType or tok.tType == nlTokType.EOL)