Fix reading RGB and CMYK IPTC images

[radarhere]

When opening IPTC images at the moment, self._mode might be set to a single channel from RGB or CMYK.

Pillow/src/PIL/IptcImagePlugin.py

Lines 109 to 112 in d80cf0e

	elif layers == 3 and component:
	self._mode = "RGB"[id]
	elif layers == 4 and component:
	self._mode = "CMYK"[id]

That's incorrect, as we don't have R, G, B, C, M, Y or K modes. Your first instinct might be to say that these should be the rawmode instead, but IptcImagePlugin just passes the image data on to be opened by another plugin. Instead, this PR uses Image.merge() to combine the image with empty bands for the other channels, to make up the full mode.

I've also made some simplifications.

Pillow/src/PIL/IptcImagePlugin.py

Lines 37 to 38 in 7b1ba29

	def _i8(c: int | bytes) -> int:
	return c if isinstance(c, int) else c[0]

is unused since 0a29d63, and so can be removed, and codec is always "iptc", so we don't have to check it.

Pillow/src/PIL/IptcImagePlugin.py

Lines 127 to 131 in d80cf0e

	ImageFile._Tile("iptc", (0, 0) + self.size, offset, compression)
	]
	def load(self) -> Image.core.PixelAccess | None:
	if len(self.tile) != 1 or self.tile[0][0] != "iptc":