Fixed reading FLI/FLC images with a prefix chunk

[twolife]

fix FLC decoder for files with prefix chunk

without the seek() the decoder continues to read bytes in the prefixchunk and have zero chance to find the 0xF1FA magic number ut needs just after

[hugovk]

Please could you add a test that fails without the fix, and passes with it?

[twolife]

This is reproducible with this sample : https://samples.ffmpeg.org/fli-flc/2422.FLC

from PIL import Image, ImageFile
ImageFile.LOAD_TRUNCATED_IMAGES = True
im = Image.open("2422.FLC")
im.seek(15)
im.save('frame15.png')

You get garbage without the patch and a clean image with the patch

[hugovk]

Can you write a unit test, with an image that can be contributed under Pillow's licence?

A unit test helps ensure this doesn't break in the future.

[twolife]

test added

[hugovk]

Thank you!