On Sunday, August 21, 2016 at 5:05:59 AM UTC+2, Terry Chia wrote:
>
> @Florian: Getting this into core Python and/or OpenSSL is a possible
> option but that will not benefit users of older Pythons (and of course
> PyPy) of which there are many.
>
Yeah, but I think that writing a faster implement
@Florian: Getting this into core Python and/or OpenSSL is a possible option
but that will not benefit users of older Pythons (and of course PyPy) of
which there are many.
@Alex: The implementation I am binding it *much* faster than the one
OpenSSL ships, even in 1.1.0 due to the strategies desc
Well fastpbkdf2 claims a 3x improvement over what Python ships. See the timeit
results in the README.
--
Aymeric.
> On 20 Aug 2016, at 22:03, Alex Gaynor wrote:
>
> AFAIK upstream Python does use an optimized PBKDF2 (and OpenSSL 1.1.0 also
> has it).
>
> Alex
>
> On Sat, Aug 20, 2016 at 4:
AFAIK upstream Python does use an optimized PBKDF2 (and OpenSSL 1.1.0 also
has it).
Alex
On Sat, Aug 20, 2016 at 4:02 PM, Florian Apolloner
wrote:
> Ups, just realized that we use pbldf2 for more than just password hashing
> -- if that where not the case the multi-library approach like the bcry
Ups, just realized that we use pbldf2 for more than just password hashing
-- if that where not the case the multi-library approach like the bcrypt
hasher uses would be preferred. That said, I would still prefer it if you
got in contact with Christian Heimes from Python to push this into Python
What was the motivation for writing your own library instead of improving
the version in python itself? Personally I do not see much gain, if you
have to install a dependency, you could as well just install bcrypt or
argon2 and ditch pbkdf2. For those people where it really makes a
difference,
Hello,
I’m +0 on adding support for this library. I also think it would be worth
trying to integrate it in Python itself so we can rely on the stdlib for the
faster version eventually.
--
Aymeric.
> On 20 Aug 2016, at 11:42, Terry Chia wrote:
>
> Greetings,
>
> 11 months ago, I opened a ti
Greetings,
11 months ago, I opened a ticket (#25395) on the bug tracker about
potentially adding a dependency on python-fastpbkdf2, a library I wrote and
maintain that provides a faster implementation of PBKDF2 than the stdlib
while maintaining API compatibility. Tim rightly pointed out that he
