Show HN: CLMM, an exercise in cryptographic minimalism

[lisper]

https://github.com/rongarret/clmm

[Fastidious]

Isn't this just similar to PGP but with a smaller public key?

[lisper → Fastidious]

Yes, this is essentially a PGP-like thing based on elliptic curves instead of RSA, so the keys and the signatures (and the code!) are all much smaller.

[rakoo → Fastidious]

Yes, minus the whole WoT part.

Note also that support for ECC is already there in the latest GnuPG, and Ed25519 is already supported for signatures. The rest is coming:

https://gnupg.org/faq/whats-new-in-2.1.html#ecc

[lisper → rakoo]

> minus the whole WoT part

CLMM is not even close to being as featureful as GnuPG. But it does a lot for being <1500 LOC (compare to >200 kLOC for GnuPG). If you want to look under the hood to see how things word, CLMM is a lot more accessible, which was its primary goal.

[cpach → lisper]

Another project which seems similar in spirit is reop: http://www.tedunangst.com/flak/post/reop

[jdjgldgskjdsjlg]

Hi, from a quick glance there seems to be a few problems that should be easy to fix:

* You should enable -Wall and -Wextra, there are a few other flags that are usually useful as well which I forget.

* You should pass the -std=c99 flag to the compiler, it fails otherwise on my setup (Debian 7.x)

Despite warnings not being enabled, I get many. Implicit declarations of strlen, asprintf, getpassn. This seems to partialy be caused by including strings.h and not string.h.

[lisper → jdjgldgskjdsjlg]

Thanks. I developed it on a Mac. I'll fix the Linux problems ASAP.

[lisper → jdjgldgskjdsjlg]

Should be fixed now.

[ecesena]

Any plan to make it tweetable as well? :)

[lisper → ecesena]

No, but it's MIT-licensed so if you want to tweet it by all means go for it.