progs/fp: testcases for GL_ARB_fragment_coord_conventions

Based on a patch from Luca Barbieri but moved the comments after
the !!ARBfp1.0 header

progs/fp/position-frc-integer.txt

@@ -0,0 +1,7 @@
+!!ARBfp1.0
+# expected: black triangle
+# brown means the wrong pixel center convention is being used
+OPTION ARB_fragment_coord_pixel_center_integer;
+MOV result.color, {0}.x;
+FRC result.color.xy, fragment.position;
+END

progs/fp/position-frc.txt

@@ -0,0 +1,6 @@
+!!ARBfp1.0
+# expected: brown triangle
+# black means the wrong pixel center convention is being used
+MOV result.color, {0}.x;
+FRC result.color.xy, fragment.position;
+END

progs/fp/position-upper-left.txt

@@ -0,0 +1,7 @@
+!!ARBfp1.0
+# expected: the yellow vertex is the bottom one
+# if it is the top one, the wrong origin convention is being used
+OPTION ARB_fragment_coord_origin_upper_left;
+MOV result.color, {0}.x;
+MUL result.color.xy, fragment.position, {.005}.x;
+END

progs/fp/position.txt

@@ -1,4 +1,6 @@
 !!ARBfp1.0
+# expected: the yellow vertex is the top one
+# if it is the bottom one, the wrong origin convention is being used
 MOV result.color, {0}.x;
 MUL result.color.xy, fragment.position, {.005}.x;
 END