On 18/08/2020 18.15, sebb wrote:
Now that the wrap_fixed=False is set, it looks like no messages are
actually changed by the rewrapping.
I think this is because the decode() method splits lines using
b'\r\n', i.e. CRLF
However it looks like all the mbox files use just LF as their line endings.
So decode will always return the body as a single chunk to
convertToWrapped() which then does nothing unless wrap_fixed=True.
Good catch! I was indeed wondering why the tests were passing on all
fronts, as one part uses the decode and the other should fail to use it
(and thus the results should be different).
Should we be changing lonely LFs to CRLF here, when flowed?
