Hi Peter, 2008/8/23 Peter Hartlich <[EMAIL PROTECTED]>: >> The only reason I can imagine is, that y < x usually
> But with mina/maxa being floating point values, why is that a problem? If y < x, then y / x < 1, whereas x / y > 1, afaik this might have effect if we use c->mina as divisor later on. > If you apply only half the patch, then in > > | else if(a < c->mina) > | h = w / c->mina; > > c->mina would have to be changed to 1.0/c->mina. Hmm, any reason why not restoring the previous h = w * c->mina instead? >> Hmm, I might have overlooked this, however I'm rather confident that >> most other WMs do it similiarly. I will add a check if minw == basew >> && minh == baseh to prevent the subtraction in this corner case. > > A window could set min and base to be the same though... This size hint > stuff is so hairy. Indeed. Kind regards, --Anselm
