On Mon, 13 Dec 2010, Harry Strongburg wrote:

>> No, there's no way to do so now. It would require a code change. I can see
>> how it could be useful.
>
> I will admit right now that I know very little about how curl
> bugs/feature requests are handled. This problem I mentioned in my
> posting is causing this bug in Transmission:
> <https://trac.transmissionbt.com/ticket/3528>. You can see comment:1 for
> the possible ways to fix this in a non-hacky way, which would require
> curl be changed. I am just wondering if there's a very little chance
> that a multiple --interface feature will be added into curl in the near
> future, and if so I can (attempt to) fix the problem in another way
> independent from curl. It's not a major problem, but it would be "nice"
> to have fixed. Thanks!

(Do note that you're asking this on the curl-users list which is about curl
the command line tool, while curl-library is the mailing list for the library
libcurl.)

We're a project with hundreds of people who have donated code at some point.
We're perhaps 10-20 persons who still donate code regularly. We have a tool
and a library that are literally feature packed.

There's always a stream of good ideas of what can be done and there's a stream
of bugs in the good things we already have implemented. And there's a limited
set of people interested in handling the individual issues in these streams.

If a suggestion is great, posted to the right place and many people think that
"hey, that's a great idea" then we easily get people interested and involved
to make it happen. So it is a good idea to make sure that you argue WHY your
suggestion is important and it is important to post to the CORRECT place. Of
course it also helps when you roll up your virtual sleeves yourself and show
that you're certainly going to do your share of the effort and not just ask
that someone else is going to spend his or her spare time to help you for
free completely one-sided.

If you check out the docs/TODO and docs/KNOWN_BUGS document you will see that
we always have a fair set of things to work on if we ever get bored. If you
hang out on the curl lists you'll see that we fix bugs at the rate of one fix
every 2-3 days. We keep busy.

Occasionally, a company pays for a feature or bug fix. But that goes outside
of the project, as then someone pays a developer to do something for (lib)curl
and the fact that someone pays somsone else doesn't change anything in how we
view the contribution, the new code, the new feature, the bug fix or whatever
it is. All contributions to the project are valued for what they bring to the
project.

From my personal view, multiple --interface sounds like a good idea but for a
feature that I don't see myself using within a reasonable time and in fact I
wouldn't even be able to try it out very good. I don't think I'll work on it
anytime soon. But of course, if someone steps up and starts working on it I
will certainly offer my help and contribute to get it done "right" for the
project. I can't speak for how likely it is for anyone else to pick it up.

(I responded with the lengthy version here so that more people get to reflect
over the situation here and how things actually work and why.)

For more numbers and stats on curl development, I blogged about this
recently:
http://daniel.haxx.se/blog/2010/09/18/curl-ten-years-of-more-code-and-contributors/

-- 
  / daniel.haxx.se - been leading this project since the dawn of mankind
-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-users
FAQ:        http://curl.haxx.se/docs/faq.html
Etiquette:  http://curl.haxx.se/mail/etiquette.html