Update pegs.nim to work at compiletime. No range errors. (#13459)

lib/pure/pegs.nim

@@ -1420,7 +1420,7 @@ type
 
   PegLexer {.inheritable.} = object ## the lexer object.
     bufpos: int                     ## the current position within the buffer
-    buf: cstring                    ## the buffer itself
+    buf: string                     ## the buffer itself
     lineNumber: int                 ## the current line number
     lineStart: int                  ## index of last line start in buffer
     colOffset: int                  ## column to add
@@ -1481,6 +1481,9 @@ proc handleHexChar(c: var PegLexer, xi: var int) =
 
 proc getEscapedChar(c: var PegLexer, tok: var Token) =
   inc(c.bufpos)
+  if c.bufpos >= len(c.buf):
+    tok.kind = tkInvalid
+    return
   case c.buf[c.bufpos]
   of 'r', 'R', 'c', 'C':
     add(tok.literal, '\c')
@@ -1508,6 +1511,9 @@ proc getEscapedChar(c: var PegLexer, tok: var Token) =
     inc(c.bufpos)
   of 'x', 'X':
     inc(c.bufpos)
+    if c.bufpos >= len(c.buf):
+      tok.kind = tkInvalid
+      return
     var xi = 0
     handleHexChar(c, xi)
     handleHexChar(c, xi)
@@ -1517,7 +1523,7 @@ proc getEscapedChar(c: var PegLexer, tok: var Token) =
     var val = ord(c.buf[c.bufpos]) - ord('0')
     inc(c.bufpos)
     var i = 1
-    while (i <= 3) and (c.buf[c.bufpos] in {'0'..'9'}):
+    while (c.bufpos < len(c.buf)) and (i <= 3) and (c.buf[c.bufpos] in {'0'..'9'}):
       val = val * 10 + ord(c.buf[c.bufpos]) - ord('0')
       inc(c.bufpos)
       inc(i)
@@ -1571,7 +1577,7 @@ proc getString(c: var PegLexer, tok: var Token) =
 
 proc getDollar(c: var PegLexer, tok: var Token) =
   var pos = c.bufpos + 1
-  if c.buf[pos] in {'0'..'9'}:
+  if pos < c.buf.len and c.buf[pos] in {'0'..'9'}:
     tok.kind = tkBackref
     tok.index = 0
     while pos < c.buf.len and c.buf[pos] in {'0'..'9'}:
@@ -1586,6 +1592,7 @@ proc getCharSet(c: var PegLexer, tok: var Token) =
   tok.charset = {}
   var pos = c.bufpos + 1
   var caret = false
+  if pos < c.buf.len:
     if c.buf[pos] == '^':
       inc(pos)
       caret = true
@@ -1661,6 +1668,13 @@ proc getTok(c: var PegLexer, tok: var Token) =
   setLen(tok.literal, 0)
   skip(c)
 
+  if c.bufpos >= c.buf.len:
+    tok.kind = tkEof
+    tok.literal = "[EOF]"
+    add(tok.literal, '\0')
+    inc(c.bufpos)
+    return
+
   case c.buf[c.bufpos]
   of '{':
     inc(c.bufpos)
@@ -1700,6 +1714,8 @@ proc getTok(c: var PegLexer, tok: var Token) =
   of '$': getDollar(c, tok)
   of 'a'..'z', 'A'..'Z', '\128'..'\255':
     getSymbol(c, tok)
+    if c.bufpos >= c.buf.len:
+      return
     if c.buf[c.bufpos] in {'\'', '"'} or
         c.buf[c.bufpos] == '$' and c.bufpos+1 < c.buf.len and
         c.buf[c.bufpos+1] in {'0'..'9'}:
@@ -1768,7 +1784,9 @@ proc arrowIsNextTok(c: PegLexer): bool =
   # the only look ahead we need
   var pos = c.bufpos
   while pos < c.buf.len and c.buf[pos] in {'\t', ' '}: inc(pos)
-  result = c.buf[pos] == '<' and (pos+1 < c.buf.len) and c.buf[pos+1] == '-'
+  if pos+1 >= c.buf.len:
+    return
+  result = c.buf[pos] == '<' and c.buf[pos+1] == '-'
 
 # ----------------------------- parser ----------------------------------------
 
@@ -2038,6 +2056,7 @@ proc escapePeg*(s: string): string =
   if inQuote: result.add('\'')
 
 when isMainModule:
+  proc pegsTest() =
     assert escapePeg("abc''def'") == r"'abc'\x27\x27'def'\x27"
     assert match("(a b c)", peg"'(' @ ')'")
     assert match("W_HI_Le", peg"\y 'while'")
@@ -2158,7 +2177,7 @@ when isMainModule:
     assert(str.find(empty_test) == 0)
     assert(str.match(empty_test))
 
-  proc handleMatches*(m: int, n: int, c: openArray[string]): string =
+    proc handleMatches(m: int, n: int, c: openArray[string]): string =
       result = ""
 
       if m > 0:
@@ -2176,3 +2195,6 @@ when isMainModule:
 
     doAssert "test1".match(peg"""{@}$""")
     doAssert "test2".match(peg"""{(!$ .)*} $""")
+  pegsTest()
+  static:
+    pegsTest()