Some corrections for GCC and extra API

[shekh]

No description provided.

[sw-team-release-gpsw]

All committers have signed the CLA.

[dnewman-gpsw]

positive pitch is required further down in the code.

[dnewman-gpsw]

Some of this seems okay, however the pitch sign change will not work for all sources. The new API additions are concerning, as I don't know all the ramifications. What problem are you trying to address?

[shekh]

the pitch sign change will not work for all sources

Can you list these sources? I tried to avoid copying the buffer when possible.
However it seems the encoder expects RG24 and BGRA in bottom-up memory order, while other formats in top-down order. This is quite inconvenient and with experiment I found the "disable flip" just works for me.

Do you expect me to undo this commit or can you simply skip it?
I don't mind if you take only useful changes in any way that works (dropping this pull request).

The API additions
GetColorFlags: I wanted to know whether YUV format is 709 or 601. I found no other way for this.
GetInputFormat: this is related to #4.
In the decoder wrapper I have this code:

    switch(format){
    case CFHD_ENCODED_FORMAT_RGB_444:
    case CFHD_ENCODED_FORMAT_RGBA_4444:
      switch(input_format){
      case COLOR_FORMAT_RGB24:
      case COLOR_FORMAT_RGB32:
      case COLOR_FORMAT_RGB32_INVERTED:
      case COLOR_FORMAT_BGRA32:
      case COLOR_FORMAT_QT32:
      obj->input_depth = 8;
      break;
      case COLOR_FORMAT_R210:
      case COLOR_FORMAT_RGB10:
      case COLOR_FORMAT_RG30:
      case COLOR_FORMAT_AR10:
      case COLOR_FORMAT_AB10:
      case COLOR_FORMAT_DPX0:
      obj->input_depth = 10;
      break;
      }
    }

The end result is, if the source is 8-bit then I assume maximum value for decoded pixel is 0xFF00.
And for 10-bit source 0xFFC0. For any other source: 0xFFFF.