Daniel Stenberg wrote:

> I see a lot of code added with very little gain for the majority of
libcurl users.
> At the same time, I can see how most of this can be done pretty
similarly with little loss at a layer outside of libcurl.

I agree: since the true libcurl integration is not yet done, code might
look as duplicated. For this reason, integration could slim down
existing code. Up to now, code intermixing mainly consists in reusing
the curl_slist structure and some internal procedures and data like
base64, boundary generation and hash table.

> So if there's no additional pros or people standing up for the
greatness of this idea, I'm sorry to say that I'm against it.

I'll wait for some other reaction until next week, then if none or
negative, I'll dissociate the MIME handler from libcurl. In this case,
I'll carry on working on it as a separate library because we (my
company) really need it, but I'm afraid to say this would'nt be
opensourced (company policy).

I'm sure there are lots of SOAP users that have renounced on using
libcurl because of the lack of a MIME parser.

If some other list readers have some interest on it, this is the time to
say so !!!

Daniel, many thanks for your attention
Patrick
Received on 2008-08-07