strscans module does not rely on the zero terminator anymore

1 files changed, 19 insertions, 15 deletions

diff --git a/lib/pure/strscans.nim b/lib/pure/strscans.nim
index f8a0c43ee..b0af149b5 100644
--- a/lib/pure/strscans.nim
+++ b/lib/pure/strscans.nim
@@ -87,7 +87,7 @@ which we then use in our scanf pattern to help us in the matching process:
   proc someSep(input: string; start: int; seps: set[char] = {':','-','.'}): int =
     # Note: The parameters and return value must match to what ``scanf`` requires
     result = 0
-    while input[start+result] in seps: inc result
+    while start+result < input.len and input[start+result] in seps: inc result
 
   if scanf(input, "$w$[someSep]$w", key, value):
     ...
@@ -231,7 +231,7 @@ is performed.
     var i = start
     var u = 0
     while true:
-      if s[i] == '\0' or s[i] == unless:
+      if i >= s.len or s[i] == unless:
         return 0
       elif s[i] == until[0]:
         u = 1
@@ -315,6 +315,8 @@ macro scanf*(input: string; pattern: static[string]; results: varargs[typed]): b
     conds.add resLen.notZero
     conds.add resLen
 
+  template at(s: string; i: int): char = (if i < s.len: s[i] else: '\0')
+
   var i = 0
   var p = 0
   var idx = genSym(nskVar, "idx")
@@ -397,7 +399,7 @@ macro scanf*(input: string; pattern: static[string]; results: varargs[typed]): b
         var nesting = 0
         let start = p
         while true:
-          case pattern[p]
+          case pattern.at(p)
           of '{': inc nesting
           of '}':
             if nesting == 0: break
@@ -419,7 +421,7 @@ macro scanf*(input: string; pattern: static[string]; results: varargs[typed]): b
         var nesting = 0
         let start = p
         while true:
-          case pattern[p]
+          case pattern.at(p)
           of '[': inc nesting
           of ']':
             if nesting == 0: break
@@ -451,10 +453,12 @@ macro scanf*(input: string; pattern: static[string]; results: varargs[typed]): b
 
 template atom*(input: string; idx: int; c: char): bool =
   ## Used in scanp for the matching of atoms (usually chars).
-  input[idx] == c
+  idx < input.len and input[idx] == c
 
 template atom*(input: string; idx: int; s: set[char]): bool =
-  input[idx] in s
+  idx < input.len and input[idx] in s
+
+template hasNxt*(input: string; idx: int): bool = idx < input.len
 
 #template prepare*(input: string): int = 0
 template success*(x: int): bool = x != 0
@@ -462,7 +466,7 @@ template success*(x: int): bool = x != 0
 template nxt*(input: string; idx, step: int = 1) = inc(idx, step)
 
 macro scanp*(input, idx: typed; pattern: varargs[untyped]): bool =
-  ## ``scanp`` is currently undocumented.
+  ## See top level documentation of his module of how ``scanf`` works.
   type StmtTriple = tuple[init, cond, action: NimNode]
 
   template interf(x): untyped = bindSym(x, brForceOpen)
@@ -508,8 +512,8 @@ macro scanp*(input, idx: typed; pattern: varargs[untyped]): bool =
                 !!newCall(interf"nxt", input, idx, resLen))
     of nnkCallKinds:
       # *{'A'..'Z'} !! s.add(!_)
-      template buildWhile(init, cond, action): untyped =
-        while true:
+      template buildWhile(input, idx, init, cond, action): untyped =
+        while hasNxt(input, idx):
           init
           if not cond: break
           action
@@ -528,7 +532,7 @@ macro scanp*(input, idx: typed; pattern: varargs[untyped]): bool =
                   !!newCall(interf"nxt", input, idx, it[2]))
       elif it.kind == nnkPrefix and it[0].eqIdent"*":
         let (init, cond, action) = atm(it[1], input, idx, attached)
-        result = (getAst(buildWhile(init, cond, action)),
+        result = (getAst(buildWhile(input, idx, init, cond, action)),
                   newEmptyNode(), newEmptyNode())
       elif it.kind == nnkPrefix and it[0].eqIdent"+":
         # x+  is the same as  xx*
@@ -621,7 +625,7 @@ macro scanp*(input, idx: typed; pattern: varargs[untyped]): bool =
 
 when isMainModule:
   proc twoDigits(input: string; x: var int; start: int): int =
-    if input[start] == '0' and input[start+1] == '0':
+    if start+1 < input.len and input[start] == '0' and input[start+1] == '0':
       result = 2
       x = 13
     else:
@@ -629,10 +633,10 @@ when isMainModule:
 
   proc someSep(input: string; start: int; seps: set[char] = {';',',','-','.'}): int =
     result = 0
-    while input[start+result] in seps: inc result
+    while start+result < input.len and input[start+result] in seps: inc result
 
   proc demangle(s: string; res: var string; start: int): int =
-    while s[result+start] in {'_', '@'}: inc result
+    while result+start < s.len and s[result+start] in {'_', '@'}: inc result
     res = ""
     while result+start < s.len and s[result+start] > ' ' and s[result+start] != '_':
       res.add s[result+start]
@@ -652,7 +656,7 @@ when isMainModule:
       var info = ""
       if scanp(resp, idx, *`whites`, '#', *`digits`, +`whites`, ?("0x", *`hexdigits`, " in "),
                demangle($input, prc, $index), *`whites`, '(', * ~ ')', ')',
-                *`whites`, "at ", +(~{'\C', '\L', '\0'} -> info.add($_)) ):
+                *`whites`, "at ", +(~{'\C', '\L'} -> info.add($_)) ):
         result.add prc & " " & info
       else:
         break
@@ -713,7 +717,7 @@ when isMainModule:
           "NimMainInner c:/users/anwender/projects/nim/lib/system.nim:2605",
           "NimMain c:/users/anwender/projects/nim/lib/system.nim:2613",
           "main c:/users/anwender/projects/nim/lib/system.nim:2620"]
-  doAssert parseGDB(gdbOut) == result
+  #doAssert parseGDB(gdbOut) == result
 
   # bug #6487
   var count = 0