| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fixes #182.
Reasoning for the change was:
I think we are just lying to ourselves when we keep PGP module in the
repository.
1. It is not used. Over a year, there is ``DeprecationWarning`` asking
anybody who uses the module to contact me. Nobody ever did. Even the
previous maintainer of M2Crypto [did not know how to use
it](https://stackoverflow.com/a/1042139/164233), and that was 2009.
Whole Stack Overflow is full of vague comments like “hopefully
M2Crypto.PGP will help you”, but I have not found on DuckDuckGo one
example of use of M2Crypto.PGP
2. The main point of M2Crypto is to be Python bindings for OpenSSL. And
truly most of the other M2Crypto code is just a tiny wrapper around
C functions. That is not so with M2Crypto.PGP. It is basically
reimplementation of PGP keys manipulation functions with Python with
rather large set of Python objects and complicated computations. There
are very few calls to any C code at all and most work is done by series
of Python functions. And, obviously, PGP != OpenSSL.
3. There is very tiny test suite for M2Crypto.PGP. It is basically just
running ``PGP.load_pubring()``, which does exercise a lot of the
module, but certainly it is far from the rigorous unit testing.
4. And of course, it is completely broken under Python 3, and I have no
idea how to fix it (especially given 1. I don't want to spend too
much time on it).
|
