[erlang-questions] Libsodium bindings

[Jesper Louis Andersen]

Hi,

jlouis/enacl is definitely maintained. It is just that there is not much to
do with the functionality in there because most of the base functionality
is present and I'll just have to add a couple of functions as people
request them.

The strength of enacl, at least compared to other solutions, is that it
contains a QuickCheck model for testing it does the right thing. Due to
work by Irina Guberman and Zane Beckwith (both working at Xaptum), this
model can also be run with QuickCheck mini. This was used to verify that
the code also worked on the MIPS architecture, and it can be used for other
embedded device work.

Enacl uses dirty schedulers, but doesn't do so for very small quick calls
under a threshold, where the move to/from the dirty scheduler is too
expensive. There is a blog post on the subject [0] where the basic idea is
this: Once the overhead of moving to a dirty scheduler is less than 5-10%
of the runtime, we just move the work to the DS. By experimenting with
dtrace, we can find that threshold point.

[0] https://medium.com/@jlouis666/erlang-dirty-scheduler-overhead-6e1219dcc7


On Thu, Dec 14, 2017 at 11:29 AM Krzysztof Jurewicz <
krzysztof.jurewicz@REDACTED> wrote:

> > Has anyone got any experience with what would be the most reasonably
> written and maintained Erlang libsodium library for cryptography?
>
> In Ercoin I’ve used https://github.com/potatosalad/erlang-libsodium.git,
> perhaps because it seemed to be the simplest one to use and least
> problematic for this particular (non-complex) application. Generally it
> looks like https://github.com/jlouis/enacl.git is the most maintained
> binding.
> _______________________________________________
> erlang-questions mailing list
> erlang-questions@REDACTED
> http://erlang.org/mailman/listinfo/erlang-questions
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20171214/a783f0e5/attachment.htm>