As suggested, I stress-tested the grid of results with a lot of scrolling around, expanding/folding groups and switching sections, and I couldn’t observe visible regressions. The performance is good too, no visible slow down.
This fixes bug #703588 too, comment #6 provides a procedure to reliably reproduce.
I’ll now try to understand the fix code-wise; at a first glance it looks ok.
As suggested, I stress-tested the grid of results with a lot of scrolling around, expanding/folding groups and switching sections, and I couldn’t observe visible regressions. The performance is good too, no visible slow down.
This fixes bug #703588 too, comment #6 provides a procedure to reliably reproduce.
I’ll now try to understand the fix code-wise; at a first glance it looks ok.