> As we discussed on IRC please comment why you use flush() on the temporary
> file so someone doesn't inadvertantly "fix" it.
I added a comment explaining the apparent oddity.
> We also discussed how our test method encode_file doesn't work in the manner
> we'd use in real life so it may be better to rewrite it to use the UNIX
> 'base64' command. We used to suggest 'uuencode' which didn't work with our
> decoding. So I do think we should test the method we suggest people use.
Yep, we should have a test like that. If we don't remove the need for base64
encoding alltogether, I'll add one in another branch.
> As we discussed on IRC please comment why you use flush() on the temporary
> file so someone doesn't inadvertantly "fix" it.
I added a comment explaining the apparent oddity.
> We also discussed how our test method encode_file doesn't work in the manner
> we'd use in real life so it may be better to rewrite it to use the UNIX
> 'base64' command. We used to suggest 'uuencode' which didn't work with our
> decoding. So I do think we should test the method we suggest people use.
Yep, we should have a test like that. If we don't remove the need for base64
encoding alltogether, I'll add one in another branch.