Since there's no reply about where ZIP files with appended 'garbage'
might come from, should this be closed?  I agree with amk; it doesn't
seem wise to ignore incorrectly formatted files without a good reason.

On a related note, it seems that correctly formatted ZIP files with
comments longer than 4k will not be recognized as a ZIP file by the
zipfile module.  There's some comment about not looking at more than the
last 4k of the file for "efficiency" reasons, but failing to interpret a
correct file seems to override the efficiency concern of
reading/searching another 60k of data.  I'll open a separate tracker
item for this, since it's not strictly related to the original issue.