I do not properly comprehend why anyone is asking for libretro. It seems to be a very young library, with a minimal userbase, and goodness knows what its CPU overhead would be. THe main website is a huge list of enumators that use it, with essentially zero technical literature on the library, and I get the sensation that the one or two users asking us to port to it are involved in its development, rather than interested in what is best for the future of ZC.
Shifting to a new library with dodgy support and a focus on flashy advertisements over practical documentation, does not seem wise.
It would probably be 2.012375% of the work though.
------------
I also see that many of the applications of libretro are for 3D rendering, which we will never, ever, need. The same user who had previously asked for us to use it, indicated a desire for a 3D Zelda game engine. That is neither the purpose of ZC, nor anything that I would ever be interested in developing, or frankly, using. The remainder are mostly console emulators. While I can see the parallel between ZC and an emulator, the internal difference is nothing close. Clearly, the user requesting this has seen SNES9x and other emulators running on libretro, and feels that ZC is similar enough that it should run on it.
More concerning, is that the library seemingly has extremely limited support for a variety of gamepads. I gathered this bit of insight from the Wiki for it. I will admit that I have not investigated it in-depth, but I still see no specific advantage to using it, especially over proven libraries that are more versatile.
Above all else, I do not have the time required to learn yet-another-API, nor do I want 2.60 to take an extra 3 to 5 years to develop because one or two people want us to change the gfx library. While that is a good idea, eventually, as allegro 4 is outdated, doing it now is simply insane.