New struct for image mode data

[Yay295]

Currently all of the data about the format/mode of an image is stored in the image itself.

https://github.com/python-pillow/Pillow/blob/ad7be550aa539dc4c79da928dad071cf0150b731/src/libImaging/Imaging.h#L80-L107

I think it would be cleaner if this information was stored in a separate struct.

struct ImageMode {
    char mode[IMAGING_MODE_LENGTH]; /* Band names ("1", "L", "P", "RGB", "RGBA", "CMYK", "YCbCr", "BGR;xy") */
    int type;                       /* Data type (IMAGING_TYPE_*) */
    int depth;                      /* Depth (ignored in this version) */
    int bands;                      /* Number of bands (1, 2, 3, or 4) */
    int pixelsize;                  /* Size of a pixel, in bytes (1, 2 or 4) */
};

Furthermore, I think some more changes could be made that would simplify coding for some calculations, and allow for more image formats to be processed (#1888).

struct BandDataType {
    char dataType; /* 'u' = unsigned integer, 's' = signed integer, 'f' = floating point */
    UINT8 numBytes;
} BAND_DATA_TYPES[] = {
    {'u', 1}, /* UINT8 */
    {'u', 2}, /* UINT16 */
    {'s', 4}, /* INT32 */
    {'f', 4}, /* FLOAT32 */
};
/* Normal bands are treated normally */
/* Alpha bands may be treated specially by some calculations */
/* Extra bands are ignored (this could be used for padding, or maybe storing non-image data) */
enum BandType { NORMAL, ALPHA, EXTRA };
struct Band {
    char *name; /* "R", "G", "B", "Y", "Cb", "Cr", etc. */
    BandDataType dataType;
    BandType type;
};
struct ImageFormat {
    char *name;      /* Format name ("1", "L", "P", "RGB", "RGBA", "CMYK", "YCbCr", "BGR;xy") */
    Band *bands;     /* Array of bands in this format */
    UINT32 numBands; /* Number of bands in this format */
    UINT32 numBytes; /* Total size in bytes of one pixel in this format */
};