On 10 October 2013 23:51, Matthew Mondor <mm_lists@pulsar-zone.net> wrote:
On Mon, 7 Oct 2013 10:54:01 +0200
Juan Jose Garcia-Ripoll <juanjose.garciaripoll@gmail.com> wrote:
> * I am not going to change ECL's license. LGPL3's restrictions on web
> applications seem stupid to me and, as experience has shown, making such a
> move will only make things worse. Already LGPL2 is a hindrance, but I can
> live with it.

This is something I'd like to understand better: what restrictions does
LGPL3 add on network services?  Does it have to do with the fact that
web applications disclose source, versus my example of using a web
server compiled using ECL?

Some here might already know my preference for MIT/BSD licenses, but in
this case even if ECL was relicensed I don't see how it could solve the
dependency on LGPL3 GMP/MPIR anyway...

I have no experience with GMP or similar libraries, but would switching to a BSD/MIT licensed library solve this problem?  I see there are a couple mentioned here (although maybe OpenSSL doesn't count):


http://crd-legacy.lbl.gov/~dhbailey/mpdist/ (although the license for this one appears to want you to fill in your contact details, despite looking like a BSD license?)

Of course BSDMP explicitly says that the API is not compatible with GMP, and the others are also probably not compatible.

> * This said, GMP v5 is insufficient for several platforms but I will
> maintain it as it is. *On platforms where GMP becomes obsolete, it will
> shift to building with "C"* (i.e. no optimized assembly code). I tested
> this on Cygwin/64 and it works -- *indeed it is part of the source tree
> right now*. If you need a better GMP, build ECL with the one that your
> operating system provides and be tied to its license.

This seems like a good compromise to me.

Michael Wood <esiotrot@gmail.com>