implicit auto return type for inline iterators

compiler/semtypes.nim

@@ -849,9 +849,13 @@ proc semProcTypeNode(c: PContext, n, genericParams: PNode,
       addParamOrResult(c, arg, kind)
       if gCmd == cmdPretty: checkDef(a.sons[j], arg)
 
-
+  var r: PType
   if n.sons[0].kind != nkEmpty:
-    var r = semTypeNode(c, n.sons[0], nil)
+    r = semTypeNode(c, n.sons[0], nil)
+  elif kind == skIterator:
+    r = newTypeS(tyAnything, c)
+  
+  if r != nil:
     # turn explicit 'void' return type into 'nil' because the rest of the 
     # compiler only checks for 'nil':
     if skipTypes(r, {tyGenericInst}).kind != tyEmpty:

doc/manual.txt

@@ -2822,7 +2822,6 @@ as there are components in the tuple. The i'th iteration variable's type is
 the type of the i'th component. In other words, implicit tuple unpacking in a 
 for loop context is supported.
 
-
 Implict items/pairs invocations
 -------------------------------
 
@@ -2847,10 +2846,11 @@ First class iterators
 There are 2 kinds of iterators in Nimrod: *inline* and *closure* iterators.
 An `inline iterator`:idx: is an iterator that's always inlined by the compiler 
 leading to zero overhead for the abstraction, but may result in a heavy
-increase in code size. Inline iterators are second class
-citizens; one cannot pass them around like first class procs.
+increase in code size. Inline iterators are second class citizens;
+They can be passed as parameters only to other inlining code facilities like
+templates, macros and other inline iterators.
 
-In contrast to that, a `closure iterator`:idx: can be passed around:
+In contrast to that, a `closure iterator`:idx: can be passed around more freely:
 
 .. code-block:: nimrod
   iterator count0(): int {.closure.} =
@@ -2873,9 +2873,7 @@ Closure iterators have other restrictions than inline iterators:
 1. ``yield`` in a closure iterator can not occur in a ``try`` statement.
 2. For now, a closure iterator cannot be evaluated at compile time.
 3. ``return`` is allowed in a closure iterator (but rarely useful).
-4. Since closure iterators can be used as a collaborative tasking
-   system, ``void`` is a valid return type for them.
-5. Both inline and closure iterators cannot be recursive.
+4. Both inline and closure iterators cannot be recursive.
 
 Iterators that are neither marked ``{.closure.}`` nor ``{.inline.}`` explicitly
 default to being inline, but that this may change in future versions of the
@@ -2937,6 +2935,14 @@ parameters of an outer factory proc:
   for f in foo():
     echo f
 
+Implicit return type
+--------------------
+
+Since inline interators must always produce values that will be consumed in
+a for loop, the compiler will implicity use the ``auto`` return type if no
+type is given by the user. In contrast, since closure iterators can be used
+as a collaborative tasking system, ``void`` is a valid return type for them.
+
 
 Type sections
 =============

tests/iter/tchainediterators.nim

@@ -12,7 +12,7 @@ iterator gaz(it: iterator{.inline.}): type(it) =
   for x in it:
     yield x*2
 
-iterator baz(it: iterator{.inline.}): auto =
+iterator baz(it: iterator{.inline.}) =
   for x in gaz(it):
     yield x*2