[FFmpeg-devel] PATCH: gdigrab work for DPI in windows

[wm4]

On Wed, 23 Sep 2015 12:04:43 -0600
Roger Pack <rogerdpack2 at gmail.com> wrote:

> From 6a972dda58bd5ab31524cd4e5326b4bcdeaeaa8c Mon Sep 17 00:00:00 2001
> From: rogerdpack <rogerpack2005 at gmail.com>
> Date: Wed, 23 Sep 2015 12:03:27 -0600
> Subject: [PATCH] gdigrab: grab right desktop size if DPI in use, based on
>  patch from Alexander Brotzge
> 
> Signed-off-by: rogerdpack <rogerpack2005 at gmail.com>
> ---
>  libavdevice/gdigrab.c | 44 +++++++++++++++++++++++++++++---------------
>  1 file changed, 29 insertions(+), 15 deletions(-)
> 
> diff --git a/libavdevice/gdigrab.c b/libavdevice/gdigrab.c
> index 9a185d4..b0faf45 100644
> --- a/libavdevice/gdigrab.c
> +++ b/libavdevice/gdigrab.c
> @@ -235,6 +235,9 @@ gdigrab_read_header(AVFormatContext *s1)
>      AVStream   *st       = NULL;
>  
>      int bpp;
> +    int vertres;
> +    int desktopvertres;
> +    float scale;
>      RECT virtual_rect;
>      RECT clip_rect;
>      BITMAP bmp;
> @@ -263,14 +266,34 @@ gdigrab_read_header(AVFormatContext *s1)
>          goto error;
>      }
>  
> -    if (hwnd) {
> -        GetClientRect(hwnd, &virtual_rect);
> -    } else {
> +    /* This will get the device context for the selected window, or if
> +     * none, the primary screen */
> +    source_hdc = GetDC(hwnd);
> +    if (!source_hdc) {
> +        WIN32_API_ERROR("Couldn't get window device context");
> +        ret = AVERROR(EIO);
> +        goto error;
> +    }
> +    bpp = GetDeviceCaps(source_hdc, BITSPIXEL);
> +
> +    scale = 1.0;
> +    if (hwnd == NULL) {
> +      /* desktop -- get the right height and width for scaling DPI */
> +      vertres = GetDeviceCaps(source_hdc, VERTRES);
> +      desktopvertres = GetDeviceCaps(source_hdc, DESKTOPVERTRES);
> +      scale = (float) desktopvertres / (float) vertres;
> +    }
> +
> +     if (hwnd) {
> +         GetClientRect(hwnd, &virtual_rect);
> +         virtual_rect.right = virtual_rect.right * scale;
> +         virtual_rect.bottom = virtual_rect.bottom * scale;
> +     } else {
>          virtual_rect.left = GetSystemMetrics(SM_XVIRTUALSCREEN);
>          virtual_rect.top = GetSystemMetrics(SM_YVIRTUALSCREEN);
> -        virtual_rect.right = virtual_rect.left + GetSystemMetrics(SM_CXVIRTUALSCREEN);
> -        virtual_rect.bottom = virtual_rect.top + GetSystemMetrics(SM_CYVIRTUALSCREEN);
> -    }
> +        virtual_rect.right = (virtual_rect.left + GetSystemMetrics(SM_CXVIRTUALSCREEN)) * scale;
> +        virtual_rect.bottom = (virtual_rect.top + GetSystemMetrics(SM_CYVIRTUALSCREEN)) * scale;
> +     }
>  
>      /* If no width or height set, use full screen/window area */
>      if (!gdigrab->width || !gdigrab->height) {
> @@ -299,15 +322,6 @@ gdigrab_read_header(AVFormatContext *s1)
>              goto error;
>      }
>  
> -    /* This will get the device context for the selected window, or if
> -     * none, the primary screen */
> -    source_hdc = GetDC(hwnd);
> -    if (!source_hdc) {
> -        WIN32_API_ERROR("Couldn't get window device context");
> -        ret = AVERROR(EIO);
> -        goto error;
> -    }
> -    bpp = GetDeviceCaps(source_hdc, BITSPIXEL);
>  
>      if (name) {
>          av_log(s1, AV_LOG_INFO,

I don't know the code, but IMO this is all kinds of fishy. Why are
there apparently two units involved, one scaled and one not? If it's
using always the same API (classic gdi/user stuff), why would it be
scaled in some cases?

Won't you run into rounding errors by using floats?