`smart_gaps` and `smart_borders` problems

[dawsers]

From this Reddit post, there were some comments about this feature missing in scroll.

I added the feature in 24872d9, but apparently there are still problems, as mentioned in the same thread:

smart_borders and smart_gaps both don't work. smart_gaps just hides all gaps when the window is the only
one in the workspace, similar to how smart_borders just hides all borders when the window is the only one in
the workspace. Also, even when default_border none, gaps inner 0 and gaps outer 0 are set, there is still some
kind of space between the screen edge and the window that makes the pointer not be able to act upon the
window contents. This behaviour is also not found in sway where if there are no gaps or borders, the window
contents extend to the edge of the screen. For example, put your pointer at the extreme edge of the screen,
open a web browser and try to scroll; this works in sway but not in scroll. Tested this on the new 1.12.10 as well.

This issue will track the problems.

[dawsers]

I haven't been able to see any problems with the current implementation, aside from scrolling layout specific differences.

From the manual:

*smart_borders* on|no_gaps|off
	If smart_borders are _on_, borders will only be enabled if the workspace
	has more than one visible child. If smart_borders is set to _no_gaps_,
	borders will only be enabled if the workspace has more than one visible
	child and gaps equal to zero.

*smart_gaps* on|off|toggle|inverse_outer
	If smart_gaps are _on_ gaps will only be enabled if a workspace has more
	than one child. If smart_gaps are _inverse_outer_ outer gaps will only
	be enabled if a workspace has exactly one child.

The current behavior matches what sway/scroll's manual says. The main difference is sway's first visible child is always maximized, while scroll's isn't by default. You can also change this in scroll by setting maximize_if_single true (default is false).

The main difference is in the command hide_edge_borders. In sway, all the windows are visible in the viewport, so the command basically hides some borders depending on the location of the window. This can only change for a window, it it is moved. Instead, scroll uses a scrolling layout, so even if a window doesn't really "move" (relative to other windows), changing focus may displace it outside of the viewport. I took the liberty of reinterpreting the command. From scroll's manual:

*hide_edge_borders* none|vertical|horizontal|both
	Hides window borders on the boundary of the tiled containers. If enabled,
	on the left-most container, the left border will be hidden, on the right-most
	container, the right border, and so on.	Default is _none_.

So, enabling this, the borders will be removed from the selected boundaries, this way you know which window is "inside" and which one is at a boundary (comparable to having a viewport that covers all the windows, even when in reality, some may be outside because of focus).

If anyone thinks there are still problems with smart_gaps and/or smart_borders, please feel free to re-open the issue or leave a comment here. Thank you.