Correct writing buffer to file in PyEncoder encode_to_file()

[radarhere]

Resolves #10096

PyEncoder.encode_to_file() is meant to mirror the behaviour of our C encoders.

In C's _encode_to_file(), we have

Pillow/src/encode.c

Lines 201 to 204 in 9712fff

	status = encoder->encode(encoder->im, &encoder->state, buf, bufsize);
	if (status > 0) {
	if (write(fh, buf, status) < 0) {

In Python, we have

Pillow/src/PIL/ImageFile.py

Lines 933 to 935 in 9712fff

	status, errcode, buf = self.encode(bufsize)
	if status > 0:
	os.write(fh, buf[status:])

There seems to have been a fairly basic mistake here in Python - we should be writing status number of bytes, not removing status number of bytes before writing. buf[:status], not buf[status:].

[akx]

status seems a misnomer?

Suggested change

	status, errcode, buf = self.encode(bufsize)
	if status > 0:
	os.write(fh, buf[status:])
	os.write(fh, buf[:status])
	n_bytes, errcode, buf = self.encode(bufsize)
	if n_bytes > 0:
	os.write(fh, buf[:n_bytes])

[radarhere]

I've pushed bytes_consumed after https://pillow.readthedocs.io/en/stable/handbook/writing-your-own-image-plugin.html#transforming

It is the codec’s responsibility to pull as much data as possible out of the buffer and return the number of bytes consumed.