The solution proposed by David Black won't eliminate all artefacts related to the thumbnail window previews. To see this, apply the proposed solution and then just type "wish" in a MATE-Terminal. The resulting window has a geometry of 200x200, and the corresponding thumbnail still gives rise to the reported artefact. Same problem with any other top-level window having a sufficiently small size.
Alternative solution: Change org.compiz.profiles.mate.plugins.thumbnail.border from the value 12 to 19 (or higher).
With this patch, I could no longer reproduce that annoying thumbnail artefact, regardless of whether the change proposed by David Black was in effect or not.
BTW: IMHO, the shadow effect generated by the original offset value of 6 (for active-shadow-x-offset, active-shadow-y-offset, inactive-shadow-x-offset and inactive-shadow-y-offset) looks better than the one corresponding to the proposed offset value of 2). Of course, this is just matter of taste.
The solution proposed by David Black won't eliminate all artefacts related to the thumbnail window previews. To see this, apply the proposed solution and then just type "wish" in a MATE-Terminal. The resulting window has a geometry of 200x200, and the corresponding thumbnail still gives rise to the reported artefact. Same problem with any other top-level window having a sufficiently small size.
Alternative solution: Change org.compiz. profiles. mate.plugins. thumbnail. border from the value 12 to 19 (or higher).
With this patch, I could no longer reproduce that annoying thumbnail artefact, regardless of whether the change proposed by David Black was in effect or not.
BTW: IMHO, the shadow effect generated by the original offset value of 6 (for active- shadow- x-offset, active- shadow- y-offset, inactive- shadow- x-offset and inactive- shadow- y-offset) looks better than the one corresponding to the proposed offset value of 2). Of course, this is just matter of taste.