i tried to solve that problem on Linux, with a wm that tiled set-sized windows in a floating grid, but it was always very, very hacky. (For the curious, github.com/halfwit/hwwm)
On Tue., Apr. 16, 2019, 9:59 p.m. Lucio De Re, <lucio.d...@gmail.com> wrote: > On 4/16/19, Marshall Conover <marzhal...@gmail.com> wrote: > > [ ... ] > > As an aside, Lucio, I'd second Ethan in that it's probably worth taking a > > look; I'd be surprised if there was more actual code to change than there > > was just ramp-up time to understand what you need to change, and a > > one-or-two hour excursion into the code would probably get you how much > > ramp-up time you need, at which point you could probably make the final > > call on whether to move forward. > > > Thanks to both of you for the hint. My problem, spelled out neatly > above, is that I have no idea how one targets the correct "window" in > X. I have completely swallowed the Plan 9 kool-aid of fine-grained > namespace separation and can't contemplate how xnest gets it wrong. > > By the same token, incidentally, I've been wondering what the X > equivalent of 9's window could be: nowhere in X's user space have I > found a way to spawn a task in a window that matches the geometry I > specify, unless that task includes the geometry among the command line > arguments. That bit of philosophy shows so clearly how different Plan > 9 and Unix really are. > > I will look, however. Frankly, if rio could provide a "little" more > support, one may be able to run firefox (I'm assuming chrome/chromium > isn't quite as liberated) in a rio window, but my efforts a while ago > flopped completely - for which part of me is greatly relieved: somehow > a firefox window on a Plan 9 background would be a constant thorn in > my side. > > Lucio. > > PS: I did look at rio, not that long ago. But I think the problem > extends to individual graphic commands. I'll see if I can get catclock > to behave itself, next. Or shove the xnext magic into a rio option. > >