fetch(1) vs browser download fails comparison

[Lowell Gilbert]

Kenneth W Cochran <kwc at TheWorld.com> writes:

> How might a file retrieved/downloaded by fetch(1) differ from the
> same file downloaded by a web browser (in this case mozilla &
> firefox)?  Specifically, if I download the free version of
> ZoneAlarm for Windows (about 10mb) by each of these methods, I
> get the file(s) and they're both the same length, but they fail
> comparison, both by cmp(1) and by differing md5 hashes.  As a
> counterexample, fetching OpenOffice (some 80mb or so) works fine
> & hash-verifies correctly.  This same behavior is with both
> 4.10-stable and 5.4-release.  Ideas?

Well, one of the differing downloads is wrong and the other is right.
Figuring out which is which would at least tell you which program is
downloading things incorrectly.  Also look at the download URLs; the
download that works correctly either way is probably using a different
protocol than the one that is incorrect, and it would be useful to
know which one is having problems.