* Add ImageMagick 7 compatibility
A new header, `imclude/MagickUtilities.h`, is created to hold the
compatibility `#define`s.
The image-conversion code in `src/Frame.cpp` received the only
major changes — instead of doing the export by hand (and having
to account for changes in the underlying API), it uses the
`MagickCore::ExportImagePixels()` function which does basically
the same work, but accounts for all of the API changes for us.
The API of that function is _unchanged_ from IM6 to IM7.
TODO: `MagickCore::ExportImagePixels()` will return an `exception`
struct if it encounters any problems. Currently the code ignores
that, which it should not.
* Add ImageMagick 7 compatibility
A new header, `imclude/MagickUtilities.h`, is created to hold the
compatibility `#define`s.
The image-conversion code in `src/Frame.cpp` received the only
major changes — instead of doing the export by hand (and having
to account for changes in the underlying API), it uses the
`MagickCore::ExportImagePixels()` function which does basically
the same work, but accounts for all of the API changes for us.
The API of that function is _unchanged_ from IM6 to IM7.
TODO: `MagickCore::ExportImagePixels()` will return an `exception`
struct if it encounters any problems. Currently the code ignores
that, which it should not.
Thanks @ferdnyc
- each class is now responsible to free whatever it allocates
- all tests passed on my machine with ffmpeg 3.2
- Clip is now more careful about freeing a reader if it allocated it as well
Removing the decode setting makes hardware supported decode break.
There must be some hidden dependency on that variable somewhere
which might also be responsible for the problems with nVidia on
Linux. TODO Remove the dependency
