Parsing gpg output for signature verification is not recommended,
as it can break easily (and doesn't work with i18n).
This patch makes use of gpgv, as suggested by gpg authors:
http://lists.gnupg.org/pipermail/gnupg-users/2004-August/023141.html
This closes #3762.
Signed-off-by: NLuca Bruno <lucab@debian.org>

Closes #2142

Most could use the each method, but because of the hack used to
disambiguate old- and new-style loops, some had to use vec::each.

(This hack will go away soon.)

Issue #1619

 - Move io, run and rand to core.
 - Remove incorrect ctypes module (use libc).
 - Remove os-specific modules for os and fs.
 - Split fs between core::path and core::os.

* Renamed str::split -> str::split_byte
* Renamed str::splitn -> str::splitn_byte
* Renamed str::split_func -> str::split
* Renamed str::split_char -> str::split_char
* Renamed str::split_chars_iter -> str::split_char_iter
* Added u8::is_ascii
* Fixed the behavior of str::split_str, so that it matches split_chars
  and split (i.e. ["", "XXX", "YYY", ""] == split_str(".XXX.YYY.", "."))
* Fixed str::split_byte and str::splitn_byte so that they handle
  splitting UTF-8 strings on a given UTF-8/ASCII byte and also handle ""
  as the others do