Optimize primop calls

We now parse function applications as a vector of arguments rather
than as a chain of binary applications, e.g. 'substring 1 2 "foo"' is
parsed as

  ExprCall { .fun = <substring>, .args = [ <1>, <2>, <"foo"> ] }

rather than

  ExprApp (ExprApp (ExprApp <substring> <1>) <2>) <"foo">

This allows primops to be called immediately (if enough arguments are
supplied) without having to allocate intermediate tPrimOpApp values.

On

  $ nix-instantiate --dry-run '<nixpkgs/nixos/release-combined.nix>' -A nixos.tests.simple.x86_64-linux

this gives a substantial performance improvement:

  user CPU time:      median =      0.9209  mean =      0.9218  stddev =      0.0073  min =      0.9086  max =      0.9340  [rejected, p=0.00000, Δ=-0.21433±0.00677]
  elapsed time:       median =      1.0585  mean =      1.0584  stddev =      0.0024  min =      1.0523  max =      1.0623  [rejected, p=0.00000, Δ=-0.20594±0.00236]

because it reduces the number of tPrimOpApp allocations from 551990 to
42534 (i.e. only small minority of primop calls are partially
applied) which in turn reduces time spent in the garbage collector.

src/libexpr/nixexpr.hh

@@ -248,6 +248,17 @@ struct ExprLambda : Expr
     COMMON_METHODS
 };
 
+struct ExprCall : Expr
+{
+    Expr * fun;
+    std::vector<Expr *> args;
+    Pos pos;
+    ExprCall(const Pos & pos, Expr * fun, std::vector<Expr *> && args)
+        : fun(fun), args(args), pos(pos)
+    { }
+    COMMON_METHODS
+};
+
 struct ExprLet : Expr
 {
     ExprAttrs * attrs;
@@ -306,7 +317,6 @@ struct ExprOpNot : Expr
         void eval(EvalState & state, Env & env, Value & v); \
     };
 
-MakeBinOp(ExprApp, "")
 MakeBinOp(ExprOpEq, "==")
 MakeBinOp(ExprOpNEq, "!=")
 MakeBinOp(ExprOpAnd, "&&")