Re: Cryptokit.Random unsuitable in cooperative multithreaded systems

[Anil Madhavapeddy <anil@xxxxxxxxxx>]

On 23 Apr 2013, at 20:35, Thomas Gazagnaire <thomas@xxxxxxxxxxxxxx> wrote:

>> From the archives: 
>> http://www.elliottoti.com/index.php?p=31
> 
> The interface is not exactly the one Daniel, David and I need but that's 
> definitely a good start (it pre-allocate strings instead of streaming). I'll 
> look into that tomorrow but Vincent, feel free to look at it if you have time.
> 

Agreed; that link was just from my last zlib tag from a few years ago.  I would 
definitely prefer something on the lines of Daniel's non-blocking codecs, as 
that makes for easier integration with network streams.

A clean implementation of the encoding is also missing (and that's where the 
parameterisation is needed to choose a time/space tradeoff, although it doesn't 
matter much at all on modern hardware).

-a