mirror of
https://github.com/brl/mutter.git
synced 2025-01-25 19:08:56 +00:00
shadow-factory: Comment fixes
This commit is contained in:
parent
b7119c55a6
commit
c2fe6a18ad
@ -541,7 +541,7 @@ blur_xspan (guchar *row,
|
|||||||
* be well predicted and there are enough different possibilities
|
* be well predicted and there are enough different possibilities
|
||||||
* that trying to write this as a series of unconditional loops
|
* that trying to write this as a series of unconditional loops
|
||||||
* is hard and not an obvious win. The main slow down here seems
|
* is hard and not an obvious win. The main slow down here seems
|
||||||
* to be the integer division for pixel; one possible optimization
|
* to be the integer division per pixel; one possible optimization
|
||||||
* would be to accumulate into two 16-bit integer buffers and
|
* would be to accumulate into two 16-bit integer buffers and
|
||||||
* only divide down after all three passes. (SSE parallel implementation
|
* only divide down after all three passes. (SSE parallel implementation
|
||||||
* of the divide step is possible.)
|
* of the divide step is possible.)
|
||||||
@ -595,7 +595,7 @@ blur_rows (cairo_region_t *convolve_region,
|
|||||||
* equally far to the left and right. If d is odd that happens
|
* equally far to the left and right. If d is odd that happens
|
||||||
* naturally, but for d even, we approximate by using a blur
|
* naturally, but for d even, we approximate by using a blur
|
||||||
* on either side and then a centered blur of size d + 1.
|
* on either side and then a centered blur of size d + 1.
|
||||||
* (techique also from the SVG specification)
|
* (technique also from the SVG specification)
|
||||||
*/
|
*/
|
||||||
if (d % 2 == 1)
|
if (d % 2 == 1)
|
||||||
{
|
{
|
||||||
|
Loading…
x
Reference in New Issue
Block a user