Hi Nick,

Actually yes, we made a relatively small expansion of the very big library 
ARM/mbedtls.

I’m not active in this community and probably they don’t know about Milagro-tls.

The plan, according with my understanding, is to reach a good level of our code 
so that anyone can safely use it and then make a pull request to mbedtls.
I think the chances of acceptance are very low since the new cipher suites that 
we have implemented aren’t still a standard and 
since our changes are relatively big in terms of a single contribution. 

Anyway making a pull request will let us known by the community there, will 
give us visibility and we can get comments and
see what to do to have a full acceptance. If the pull request will not accepted 
anyway, the plan was to to re-brand mbedtls as milagro-tls
and keep it synced with the upstream mbedtls. 

Regards,

Alessandro

>>> To whom I may concern,
>>> 
>>> I’m a developer who should take care of this repository
>>> https://github.com/apache/incubator-milagro-tls 
>>> <https://github.com/apache/incubator-milagro-tls>
>>> <https://github.com/apache/incubator-milagro-tls>.
>>> 
>>> Milagro-TLS is a project consisting in expanding the existing library
>>> https://github.com/ARMmbed/mbedtls
>>> <https://github.com/ARMmbed/mbedtls> in order to support two new
>>> pairing-based key-exchange algorithm as explained here
>>> https://datatracker.ietf.org/doc/draft-budronimccusker-milagrotls/
>>> <https://datatracker.ietf.org/doc/draft-budronimccusker-milagrotls/>.
>>> 
>>> From my point of view, instead of having an own repository as it is
>>> now, it would be better to have a fork to ARMmbed/mbedtls, so that it
>>> would be easier to maintain and it will allow us to make a pull
>>> request when the right time will come.
>> 
>> Are you saying the milagro-tls library is a fork of mbedtls
>> with relatively little change?
>> 
>> A quick look at mbedtls tells me it has what looks like a healthy
>> community quite separate from milagro.  Are you active in, or at
>> least known within, that community?
>> 
>> If you're saying what I think you are, it might make more sense
>> for you and anyone else concerned with the library to work with
>> them there, to contribute and maintain whatever enhancements are
>> needed by Milagro.  The TLS lib then becomes a prerequisite
>> rather than a component of Milagro.
>> 
>> Otherwise you'd presumably need to sync regularly, and it'll
>> make the job harder if you're not working with them.
>> 
>> Unless milagro's needs could be implemented in a modular
>> fashion to complement rather than replace mbedtls?
>> If that works then it matters much less how to proceed.
>> 
>>> It would be possible to achieve this?
>> 
>> The first question has to be, what exactly are we trying to achieve
>> in having a separate library in the first place?  Then we move on
>> to the question of how best to make it work.
>> 
>> -- 
>> Nick Kew
> 

Reply via email to