On Fri March 19 2010 15:34:54 Daniel Stenberg wrote:
> My biggest concern with adding the field to the struct is that we grow the
> struct for a very specific and narrow use-case that will be used very
> rarely. I know we have the same kind of thing on numerous places, but I
> still would like to see this kind of constructs in the code reduced rather
> than increased. It isn't only the binary size of the struct that is
> trouble-some, but the shear amount of fields within the struct that at
> times need to be understood and parsed by human eyes.

I agree.

> Putting it as an extra argument will add a local argument on the stack but
> will keep the main struct untouched.

That's one point of view. On the other hand, it has to be parsed in all of the
SSL backends separately, instead of one field at one place.

I am fine with either variant. You choose. If it the negative value based one
is the way to go, we need to decide what to return in case of EWOULDBLOCK.

Kamil
-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-library
Etiquette: http://curl.haxx.se/mail/etiquette.html
Received on 2010-03-19