Re: Cryptokit.Random unsuitable in cooperative multithreaded systems

[David Sheets <sheets@xxxxxxxxxxxx>]

On Tue, Apr 23, 2013 at 5:48 PM, Daniel Bünzli
<daniel.buenzli@xxxxxxxxxxxx> wrote:
>
>
> Le mardi, 23 avril 2013 à 15:39, Thomas Gazagnaire a écrit :
>
>> zlib is heavily used in Git, so that could be quite good to have a pure zlib 
>> implementation as well. I might look into that then ...
>
> I'd be quite interested in this aswell other reasons -- e.g. I would like to 
> have a hook in Vg to be able to compress the internal streams in PDFs. That 
> would also make things like a pure PNG codec quite easy to write.

ocaml-cohttp would benefit quite a lot from this as well as
zlib/deflate are the basis of most of the HTTP compression encodings.

> I don't know the kind of interface you have in mind but it would be nice to 
> make it independent of the asynchronous library used. I have a simple pattern 
> for that here [1], it's used both by jsonm [2] and vg [3] (and will 
> eventually be used by xmlm once I get to write version 2).
>
> Best,
>
> Daniel
>
>
> [1] https://github.com/dbuenzli/nbcodec/blob/master/RATIONALE
> [2] https://github.com/dbuenzli/jsonm/blob/master/src/jsonm.mli
> [3] https://github.com/dbuenzli/vg/blob/master/src/vg.mli#L564
>
>
>