Feel free to ask more supporting information to your bug, if the maintainers don't understand what you mean by the bug title, this bug might be closed...
(In reply to comment #1) > if the maintainers don't understand what you mean by the bug title, this bug > might be closed... I'll try to clarify: When emerging dev-libs/botan-1.10.1 with USE=ssl the botan source code wants to #include <openssl/ecdsa.h> and the botan ebuild should thus RDEPEND on dev-libs/openssl[-bindist] because dev-libs/openssl[bindist] excludes everything about ECDSA because of an unclear or risky patent situation. On systems where dev-libs/openssl is installed with USE=bindist botan fails to build with an ugly error about not finding ecdsa.h. If botan would explicitly RDEPEND on openssl[-bindist] then portage would know what the problem is and could tell the user about it before trying to emerge.
is there no way to make the ecdsa stuff optional?
(In reply to comment #3) > is there no way to make the ecdsa stuff optional? configure.py can be invoked with --disable-module=ecdsa What would be the correct thing to control that flag? USE=bindist again? I was discussing with jmbsvicetto on IRC how I think USE=bindist is missing out a lot of very nice functionality in portage, and how I would expect rather one USE flag per functionality (like usual) than one USE flag per potential legal problem. That's another debate, but maybe worth considering. In any case, fixing botan one way or another right now is important imo.
Let's give the author a few days to provide his opinion on USE=bindist vs USE=whateverhemaywant. If unresponsive I'll just go ahead and do bindist to match openssl.
Added bindist USE to botan-1.10.3 Thanks!