Do you plan to support mbedTLS?

[zhaojh329]

Do you plan to support mbedTLS?

[bl4ck5un]

Same question!

[deleisha]

Yes, we intend to support. Contributors welcome

[realquantumcookie]

Right. mbedtls is much easier to build compare to OpenSSL, and it is great to see a lib that supports SSL with libuv with embedtls. Looking forward to the new implemented embedtls version of evt-tls!

[deleisha]

Experimentation on mbedtls started, Either I am not fluent yet on it or their API are complete yet.
Using mbedtls as state black box is little difficult.

[realquantumcookie]

mbedTLS-Tutorial
This link could probably help. It describes how to create TLS connected Sockets using it.
Good Luck : )

[realquantumcookie]

I just read a bit about the tutorial. embedTLS is hard to implement into evt-tls.
It replaces the original read() function in UNIX and Windows API. Rather than using read(), it let you use its own function.
I think that implementing embedTLS support could be compicated and time-consuming since we might need to modify the implemention inside libuv.

[realquantumcookie]

As LibUV-mbedtls do, he re-wrote bio.c from mbedtls and then wrap mbedtls over libuv. I think this could be a good solution for us.

[deleisha]

Thanks @ToiletCommander for your effort.
Perhaps, we don't want to go wolfssl way for mbedtls as that will involved maintaining the bio.c. I did a feature request to mbedtls. They leave network IO(nio) handling to the user and don't do buffering.
So, we need to come up with NIO part. Now we know what need to be done.
I am working on now to have a unified error handling.

[realquantumcookie]

Just one question, as you requested the feature, when is the feature going to be implmented?
Because I am uncertain about the version of the mbedtls lib to download.

[deleisha]

I don't think it will be in the near future.
We can go ahead without that as well.
Please get the latest release version because that's the version that I am planning to checked in and use.