Add unmasked { } statement.

This reestablishes an "all on" execution mask for the gang, which can
be useful for nested parallelism..

tests/unmasked-1.ispc

@@ -0,0 +1,17 @@
+
+export uniform int width() { return programCount; }
+
+
+export void f_f(uniform float RET[], uniform float aFOO[]) {
+    float a = aFOO[programIndex]; 
+    if (a == 2) {
+        unmasked {
+            a = 0;
+        }
+    }
+    RET[programIndex] = a; 
+}
+
+export void result(uniform float RET[]) {
+    RET[programIndex] = 0;
+}

tests/unmasked-2.ispc

@@ -0,0 +1,18 @@
+
+export uniform int width() { return programCount; }
+
+
+export void f_f(uniform float RET[], uniform float aFOO[]) {
+    float a = aFOO[programIndex]; 
+    int count = 0;
+    for (int i = 0; i < a; ++i) {
+        unmasked {
+            ++count;
+        }
+    }
+    RET[programIndex] = count; 
+}
+
+export void result(uniform float RET[]) {
+    RET[programIndex] = programCount;
+}

tests/unmasked-3.ispc

@@ -0,0 +1,19 @@
+
+export uniform int width() { return programCount; }
+
+
+export void f_f(uniform float RET[], uniform float aFOO[]) {
+    float a = aFOO[programIndex]; 
+    int count = 0;
+    for (int i = 0; i < a; ++i) {
+        ccontinue;
+        unmasked {
+            ++count;
+        }
+    }
+    RET[programIndex] = count; 
+}
+
+export void result(uniform float RET[]) {
+    RET[programIndex] = 0;
+}