Handle headers with multiple encodings

[cmatsuoka]

It came to my attention that email.header.decode_header() returns a list of (string, encoding) pairs and our _charset_decoder function was handling only the first pair in this list. This was especially visible when decoding headers such as "=?Windows-1252?Q?Mauricio_Sim=F5es?= bla@bla" where the email address, which was decoded in the second pair, was lost in the final converted string.

Iterate over the list pairs and convert them separately to address this issue.

27/11 update: Additional commits deal with cosmetic changes to address pylint warnings, vim indent helpers to prevent mixing tabs and spaces, and a fix for mixed tabs and spaces.

[caiobegotti]

Shouldn't these new commits be in a separate PR? :-X

[cmatsuoka]

They should, but apparently Github shoves them into the previous one. I guess you'll have to pull selectively from the command line.

[caiobegotti]

Hmm isn't it because you are re-using an old branch for these commits instead of a new one?

[cmatsuoka]

It could be, I'm developing in master.