Thanks for sending in the sample. I took a look and found the cause. But I am not sure whether there is a way for us to handle this situation.
Basically the ID3 tag chunk that was originally in the file was corrupt. Whichever software program wrote the tag, got it wrong. It came with a chunk size that is about 85 times the actual file size. So when MC reads the file, and found the incorrect chunk size, it skips the chunk and stops reading anything beyond that point. It appears that the software that wrote the tags did not follow the fact that AIFF files are defined to be big-endian. So that huge chunk size value, I suspect, is the result of flipping the byte order of the correct chunk size.
Because of this error, when MC tried to write tags, it assumes that there is no existing ID3 chunk in the file and proceeded to write a new chunk at the end of the file. This was repeated several times. Because these new chunks have been written at the end, MC would skip them too when reading the file. If you try to let MC write tags repeatedly, the file would grow longer and longer but the tags are still not readable. We need to find a way of detecting and fixing this situation.