Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=24784 Signed-off-by: Ziqing Hui zhui@codeweavers.com --- dlls/comctl32/tests/toolbar.c | 1 - dlls/comctl32/toolbar.c | 62 ++++++++++++++++++++++++++++++++--- 2 files changed, 57 insertions(+), 6 deletions(-)
On 3/4/20 12:42 PM, Ziqing Hui wrote:
- has_alpha = FALSE;
 - if (bmp.bmBitsPixel == 32)
 - {
 for (pixel = bits; pixel < bits + cx * cy; pixel++){if ((*pixel >> 24) != 0){has_alpha = TRUE;break;}}- }
 - if (has_alpha)
 - {
 /* gray the src image */for (pixel = bits; pixel < bits + cx * cy; pixel++){DWORD alpha = (*pixel >> 24) & 0x000000ff;DWORD red = (*pixel >> 16) & 0x000000ff;DWORD green = (*pixel >> 8) & 0x000000ff;DWORD blue = (*pixel) & 0x000000ff;DWORD gray = (red * 299 + green * 587 + blue * 114) / 1000;gray = gray * alpha / 0xff;*pixel = ((alpha << 24) | (gray << 16) | (gray << 8) | gray);}bf.BlendOp = AC_SRC_OVER;bf.BlendFlags = 0;bf.SourceConstantAlpha = 0xff;bf.AlphaFormat = AC_SRC_ALPHA;GdiAlphaBlend(hdc, x, y, cx, cy, hdcImage, 0, 0, cx, cy, bf);goto done;- }
 
I think it's possible ILS_SATURATE should be used for that. Have you checked rendering on Windows with this flag?
No, I didn't check it. I don't know this flag before. I'll give it a try.
On 3/4/20 6:04 PM, Nikolay Sivov wrote:
On 3/4/20 12:42 PM, Ziqing Hui wrote:
+ has_alpha = FALSE; + if (bmp.bmBitsPixel == 32) + { + for (pixel = bits; pixel < bits + cx * cy; pixel++) + { + if ((*pixel >> 24) != 0) + { + has_alpha = TRUE; + break; + } + } + }
+ if (has_alpha) + { + /* gray the src image */ + for (pixel = bits; pixel < bits + cx * cy; pixel++) + { + DWORD alpha = (*pixel >> 24) & 0x000000ff; + DWORD red = (*pixel >> 16) & 0x000000ff; + DWORD green = (*pixel >> 8) & 0x000000ff; + DWORD blue = (*pixel) & 0x000000ff; + DWORD gray = (red * 299 + green * 587 + blue * 114) / 1000; + gray = gray * alpha / 0xff; + *pixel = ((alpha << 24) | (gray << 16) | (gray << 8) | gray); + } + bf.BlendOp = AC_SRC_OVER; + bf.BlendFlags = 0; + bf.SourceConstantAlpha = 0xff; + bf.AlphaFormat = AC_SRC_ALPHA; + GdiAlphaBlend(hdc, x, y, cx, cy, hdcImage, 0, 0, cx, cy, bf); + goto done; + }
I think it's possible ILS_SATURATE should be used for that. Have you checked rendering on Windows with this flag?
Hi Nikolay,
I tried rendering an 32-bit bitmap with ILS_SATURATE on my win10 machine, it does make the image gray as expected. So I think I should implement ILS_SATURATE for imagelist in wine first, then fix toolbar. What do you think?
Thanks,
Ziqing
On 3/4/20 6:28 PM, Ziqing Hui wrote:
No, I didn't check it. I don't know this flag before. I'll give it a try.
On 3/4/20 6:04 PM, Nikolay Sivov wrote:
On 3/4/20 12:42 PM, Ziqing Hui wrote:
+ has_alpha = FALSE; + if (bmp.bmBitsPixel == 32) + { + for (pixel = bits; pixel < bits + cx * cy; pixel++) + { + if ((*pixel >> 24) != 0) + { + has_alpha = TRUE; + break; + } + } + }
+ if (has_alpha) + { + /* gray the src image */ + for (pixel = bits; pixel < bits + cx * cy; pixel++) + { + DWORD alpha = (*pixel >> 24) & 0x000000ff; + DWORD red = (*pixel >> 16) & 0x000000ff; + DWORD green = (*pixel >> 8) & 0x000000ff; + DWORD blue = (*pixel) & 0x000000ff; + DWORD gray = (red * 299 + green * 587 + blue * 114) / 1000; + gray = gray * alpha / 0xff; + *pixel = ((alpha << 24) | (gray << 16) | (gray << 8) | gray); + } + bf.BlendOp = AC_SRC_OVER; + bf.BlendFlags = 0; + bf.SourceConstantAlpha = 0xff; + bf.AlphaFormat = AC_SRC_ALPHA; + GdiAlphaBlend(hdc, x, y, cx, cy, hdcImage, 0, 0, cx, cy, bf); + goto done; + }
I think it's possible ILS_SATURATE should be used for that. Have you checked rendering on Windows with this flag?
On 3/4/20 6:35 PM, Ziqing Hui wrote:
Hi Nikolay,
I tried rendering an 32-bit bitmap with ILS_SATURATE on my win10 machine, it does make the image gray as expected. So I think I should implement ILS_SATURATE for imagelist in wine first, then fix toolbar. What do you think?
Hi,
Sure, sounds good. Please add a test for it at imagelist level, small 4x4 / 8x8 image is enough, showing that grayscale conversion happens.
Thanks,
Ziqing
On 3/4/20 6:28 PM, Ziqing Hui wrote:
No, I didn't check it. I don't know this flag before. I'll give it a try.
On 3/4/20 6:04 PM, Nikolay Sivov wrote:
On 3/4/20 12:42 PM, Ziqing Hui wrote:
+ has_alpha = FALSE; + if (bmp.bmBitsPixel == 32) + { + for (pixel = bits; pixel < bits + cx * cy; pixel++) + { + if ((*pixel >> 24) != 0) + { + has_alpha = TRUE; + break; + } + } + }
+ if (has_alpha) + { + /* gray the src image */ + for (pixel = bits; pixel < bits + cx * cy; pixel++) + { + DWORD alpha = (*pixel >> 24) & 0x000000ff; + DWORD red = (*pixel >> 16) & 0x000000ff; + DWORD green = (*pixel >> 8) & 0x000000ff; + DWORD blue = (*pixel) & 0x000000ff; + DWORD gray = (red * 299 + green * 587 + blue * 114) / 1000; + gray = gray * alpha / 0xff; + *pixel = ((alpha << 24) | (gray << 16) | (gray << 8) | gray); + } + bf.BlendOp = AC_SRC_OVER; + bf.BlendFlags = 0; + bf.SourceConstantAlpha = 0xff; + bf.AlphaFormat = AC_SRC_ALPHA; + GdiAlphaBlend(hdc, x, y, cx, cy, hdcImage, 0, 0, cx, cy, bf); + goto done; + }
I think it's possible ILS_SATURATE should be used for that. Have you checked rendering on Windows with this flag?