Re: Cryptokit.Random unsuitable in cooperative multithreaded systems

[Daniel BÃnzli <daniel.buenzli@xxxxxxxxxxxx>]

Le mardi, 23 avril 2013 Ã 19:44, Vincent B. a Ãcrit :
> If you want, I can take this code, refactor it into a more standard
> format (like putting tests in a separate folder for example), and make  
> an OPAM/obuild (or OASIS) package out of it (and maybe try to look for  
> where I could improve performance.

I had my information theory course almost 15 years ago (gasp) so I may be a 
little bit rusty on compression. But didn't this person just implement the easy 
bits --- inflate --- IIRC in deflate you have quite a few design 
choices/strategies to make/take ?

Besides the code really looks toyish, it seems to be allocation hell in there. 
But it can make nice pseudo-code to read the spec.  

Daniel