Re: [hash2] Formal Review Begins

From: Christopher Kormanyos (e_float_at_[hidden])
Date: 2024-12-10 20:06:58

> The review of Hash2 by Peter Dimov
> and Christian Mazakas begins 

Christopher's Review.

DISCLAIMER: I know the review manager.
I highly respect Matt as a person and
through his open-source work. Furthermore,
I work in a dedicated fashion with Matt
on new, potentially upcoming Boost-like
projects. I do not receive any open-source
money. I operate as a self-funded,
independent researcher in FOSS.

Here is my review of the proposed
Boost.Hash2

- What is your evaluation of the design?

Exceptional. It is clean with terse,
well-deocumented code. The design addresses
a sorely neglected area in C/C++ programming
and makes strong, decisive strides to move
forward.

- What is your evaluation of the implementation?

It's simply fantastic. And I will qualify
this further below.

- What is your evaluation of the documentation?

It is fine and I could use the library
immediately within about 3 minutes.

- What is your evaluation of the potential usefulness
of the library? Do you already use it in industry?

The library is highly useful. At the moment,
it lacks some of the extreme quality
aspects that my industrial area would mandate.
But it is not far off from becoming a viable
secure component. I'd be willing to work
toward improving it with the authors.

What I'd like to see short-term:
 * Handle enhanced compiler warnings.
 * Include a subset of NIST testing.
 * Fuzzing run(s) on some hashes in CI.
 * I think SHA-3 is worthy of inclusion.

- Did you try to use the library? With which
compiler(s)? Did you have any problems?

Yes. I performed a whole slew of PC-based
tests with the standard vanilla compiler
crew including GCC, VC, clang in
a few versions. Given the (already)
extremely well-tested functionality
of Hash2, there were no problems.

But here comes the fun part. When I studied
the implementation it seemed highly portable.
So I put it onto several embedded bare-metal
systems. The code compiled and ran on 32-bit
microcontrollers and even ran on an 8-bit
controller. Rarely do we find such portability
and quality in Boost-proposed libraries,
or anywhere for that matter. Kudos on
a job well done! I can provide ROM-sizes
and microsecond run-time comparisons
if needed.

- How much effort did you put into your evaluation?
A glance? A quick reading? In-depth study?

An in-depth study.

- Are you knowledgeable about the problem domain?

Over the decades, my "day-job" has involved
delivery of hundreds of millions of
high-volume, safety-critical devices,
mostly having some kind of crypto in them.
I think I'm somewhat knowledgeable
in this area. I've had determining
roles including library author,
library client, security auditor and
also specified the commercial acquisition
of numerous libraries and silicon vendor
crypto-accelerators.

Ensure to explicitly include with your review:
ACCEPT, REJECT, or CONDITIONAL ACCEPT
(with acceptance conditions).

I vote to unconditionaly ACCEPT Hash2.

Thank you to the authors forthis seminal
work and bold step forward.

Thanks Matt for managing this rich
and lively review.

Kind regards, Christopher Kormanyos

   On Tuesday, December 10, 2024 at 02:56:10 PM GMT+1, Matt Borland via Boost <boost_at_[hidden]> wrote:  
 
 
> I have read over the documentation for the Hash2 library, and what follows
> are my notes.
> 

Peter, thank you for the documentation review. I always find these beneficial.

Matt
_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost