You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I get some very strange bugs when restoring multiple windows in a 4x4 grid by appending a layout and then opening terminal windows whose titles adjust are changed after initial window creation. When I use the old workaround of unmapping all the windows in the workspace, appending the layout, then remapping the windows, it works fine as it always has.
How it should look (and how it does look when I restore it in the old way):
How it looks when the windows are swallowed after a title change:
These are not mere graphical artifiacts, because if I keep pressing enter you can see that the prompt never goes below the top of the "artifact", so the window must actually be ending there, and the tiling has completely broken.
Also thanks @izzel for your work on this, it's a great improvement to have. In most situations it has worked perfectly for me, and I'm not sure if this weird bug is directly related to this change or if this change is bringing to light a bug somewhere else in the codebase.
I misunderstood the variable needs_reshape in x_reparent_window. x_reparent_window fairly assumes that the window rect won't change because it was designed with floating stickies in mind. Copying the window_rect state in x_reparent_window fixes the problem, but I worry that other data needs to be copied. @stapelberg Would child_mapped need to be copied as well? Is there anything else I'm missing in con_state that would need to be copied?
The text was updated successfully, but these errors were encountered:
I don’t see a link to logs.i3wm.org. Did you follow https://i3wm.org/docs/debugging.html? (In case you actually provided a link to a logfile, please ignore me.)
Originally posted by @JonnyHaystack in #1668 (comment)
I misunderstood the variable needs_reshape in x_reparent_window. x_reparent_window fairly assumes that the window rect won't change because it was designed with floating stickies in mind. Copying the window_rect state in x_reparent_window fixes the problem, but I worry that other data needs to be copied.
@stapelberg Would child_mapped need to be copied as well? Is there anything else I'm missing in con_state that would need to be copied?
The text was updated successfully, but these errors were encountered: