Should padding bytes be passed to glteximage2d - opengl

Question

Title itself explains the question.
The last parameter of glTexImage2D is the array of bytes (unsigned, signed depends).
Should rgb array contain padding bytes or not?

Answer 1

Should RGB array contain padding bytes or not?
That entirely depends on your needs. You can configure OpenGL to accept various data layouts. See the reference documentation of glPixelStore, the unpack parameters are what you should look at.
Padding bytes are normally found between between rows, to fill up to a certain alignment. The unpack alignment specifies the byte alignment of each row.
If your pixels are 8 bit per component, but packed into 4 bytes each with a padding byte, you can specify that, by declaring the data type to be GL_UNSIGNED_INT_8_8_8_8; if you use a type/internal type with less than 4 components the excessive bytes are ignored.

Answer 2

The short answer is no. I typically create an array of structures.
The structure would look like:
typedef struct
{
unsigned char red;
unsigned char green;
unsigned char blue;
unsigned char alpha;
} AlphaPixelBytes;
Then the array I create would look like:
AlphaPixelBytes bitmapData[NUMTEXTUREPOINTS];
You can then use bitmapData as the last argument to glTexImage2D when you create the texture or you can pass NULL as the last argument to create an empty texture you'll later populate with glTexSubImage2D.
This sort of array of structures is also useful as a data source for an NSBitmapImageRep for use (for example) in exporting a PNG file of your texture.
EDIT:
Sorry, I didn't notice you're dealing with RGB not RGBA data. The structure for RGB would look like:
typedef struct
{
unsigned char red;
unsigned char green;
unsigned char blue;
} PixelBytes;
And the array:
PixelBytes bitmapData[NUMTEXTUREPOINTS];