diff --git a/drm/rotate.vs.asm b/drm/rotate.vs.asm
index c125137..673cdaa 100644
--- a/drm/rotate.vs.asm
+++ b/drm/rotate.vs.asm
@@ -1,18 +1,17 @@
-# CONST[0] = {0.159155, 0.5, 6.283185, -3.141593}
-# CONST[1] = {rotate, _, _, _}
+-- CONST[0] = {0.159155, 0.5, 6.283185, -3.141593}
+-- CONST[1] = {rotate, _, _, _}
 
-# each instruction is only allowed to use a single unique `const` address
-#
-# instructions may use multiple `temp` addresses, so const[1] is moved to
-# temp[0]:
-#
+-- each instruction is only allowed to use a single unique `const` address
+--
+-- instructions may use multiple `temp` addresses, so const[1] is moved to
+-- temp[0]:
+--
 temp[0].x    = VE_ADD  const[1].x___  const[1].0___ ;
 
-
-# ME_SIN and ME_COS are clamped to the range -π to +π prior to the sin/cos
-# calculation.
-#
-# This 3-instruction sequence linearly remaps the range [-∞,+∞] to [-π,+π]
+-- ME_SIN and ME_COS clamp their inputs to [-π,+π] prior to the sin/cos
+-- calculation.
+--
+-- This 3-instruction sequence remaps the range [-∞,+∞] to [-π,+π]
 temp[1].x    = VE_MAD   temp[0].x___  const[0].x___  const[0].y___ ;
 temp[2].x    = VE_FRC   temp[1].x___ ;
 temp[3].x    = VE_MAD   temp[2].x___  const[0].z___  const[0].w___ ;
@@ -20,11 +19,11 @@ temp[3].x    = VE_MAD   temp[2].x___  const[0].z___  const[0].w___ ;
 temp[4].x    = ME_SIN   temp[3].___x ;
 temp[4].y    = ME_COS   temp[3].___x ;
 
-# with sin and cos calculated, this now ordinary two-dimensional rotation:
-#
-#  x = position.x * cost - position.y * sint
-#  y = position.x * sint + position.y * cost
-#  z = 0
-#  w = 1
+-- with sin and cos calculated, this now ordinary two-dimensional rotation:
+--
+--  x = position.x * cost - position.y * sint
+--  y = position.x * sint + position.y * cost
+--  z = 0
+--  w = 1
 temp[5].xy   = VE_MUL  input[0].yy__   temp[4].xy__ ;
  out[0].xyzw = VE_MAD  input[0].xx01   temp[4].yx01  temp[5].-xy00 ;