fixed issue with Content-Transfer-Encoding during encryption

decode payload before encryption

Content type not passed to encrypted submime causing bug where is only html is sent for example, the email client will not decode the html before displaying the email.

* Add information about enabling automatic
  public key download via gpg.conf

Added starttls support for SMTP

Install: automatic public key retrieval

…into master

Reviewed-on: https://git.disroot.org/Disroot/gpg-lacre/pulls/33

Provide a simple Python script that would linsten on a given port and print to
standard output any message received via SMTP on that port.

This script will then be used to automatically test gpg-mailgate with different
scenarios (unknown recipient key, RSA key, elliptic curve key).

- Use an environment variable to point at the configuration file while
  strating gpg-mailgate.py.

- Unify paths: store temporary config, logs and anything else under 'test'
  directory.

- Configure more tests (RSA, Ed25519).

- Add test descriptions to be shown before they're started.

Since it's not so easy to encrypt a message exactly the same way twice, we
only verify if the message has been encrypted or not.

Introduce minor changes to the library itself, because it doesn't work very
well with modern GnuPG.

Also, include GnuPG directory (pointed at by --homedir option).

To let the user know that tests produce logs, include a message at the end of
the test output informing about locations of E2E and Mailgate logs.

Also, extract some constants.

- Move things to configuration file where appropriate (logging format, etc.).

- Rework execute_e2e_test signature to simplify it and get rid of keyword
arguments.

- Simplify output.

- Include a header comment in configuration file.

Extract a function to calculate GPG commands to be executed and cover it with
unit tests.

Remove unused msgout files, fix addresses in msgin files.

Do not keep test/keyhome/random_seed in repository.

…s into master

Reviewed-on: https://git.disroot.org/Disroot/gpg-lacre/pulls/51
Reviewed-by: muppeth <muppeth@[email protected]>

Use lib2to3 automatic migration tool provided by Python 2.x to convert
codebase to new idioms.

Command line:

find . -type f -name '*.py' \
    -exec python2.7 -m lib2to3 \
    -f all -f idioms -f buffer -f set_literal -f ws_comma -w \
    '{}' '+'

- Use b'' (byte strings) where appropriate.

- Fix indentation.

- Replace python2.x references with python3.x.

Fix bytes sequences handling after auto-migration.

- Makefile: add 'unittest' to .PHONY targets.
- Remove unnecessary #! line from e2e_test.py.
- Extract Python path to test/e2e.ini file.

- Let the user overwrite Python binary name while calling make.

- Use environment variable set by make to instruct e2e_test.py which binary it
should call to execute Python code.

Tests kept breaking (not failing) randomly with "Broken pipe" errors.  Rework
how processes are spawned to make sure that it doesn't break them again.

- Explain how to specify Python binary path used during tests.

- Mention test logs.

…nto master

Reviewed-on: https://git.disroot.org/Disroot/gpg-lacre/pulls/58
Reviewed-by: muppeth <muppeth@[email protected]>

Migrate to Python 3.x

we need to convert to string to concatenate