I yet have to verify and test.As Vincent pointed out in 1), it's bad design. I use Fselect(1,...) because it seemed elegant from what I read, but I misunderstood the purpose at the first time and forgot to fix it after. In STinG API, this function does not exist, so you have to be at 100% and compute the timeout yourself.
So keep MiNT's behavior as it is now, I will spend this evening and tomorrow for tests and release fixes instead of working on the new Troll. I'm not happy right now, but it will pass.
PS@JFL: I don't mind, my GFA code is old, and I consider my programs as crap.
Cheers, Raj Le 06/02/2015 12:21, J. F. Lemaire a écrit :
On 6 February 2015 at 11:32, Vincent Rivière <vincent.riviere@freesbee.fr> wrote:We assumed that it was safe, because no programs were supposed to deliberately call Fselect(1,...). But > we were wrong, that change broke MyAES and Pierre's software.[...]Finally, I think we should consider to revert the Fselect() change in FreeMiNT, since it broke old software, and it is no more useful for new binaries...I'm not sure Pierre and Olivier will appreciate their current developments being called "old software" :-) Cheers, JFL