To reproduce, load the testcase in a debug build. ###!!! ASSERTION: CellIterator program error: 'PR_FALSE', file /Users/admin/trunk/mozilla/layout/tables/nsTableFrame.cpp, line 7148 I think this is related to bug 315549.
Created attachment 234590 [details] [diff] [review] wip I know close to nothing about this code so this is just a stab in the dark... Maybe we shouldn't require a cell frame unless we are painting for a cell? This patch fixes the assertion.
Created attachment 234591 [details] Screenshot (with "wip" applied) Note the missing pixel on bottom border.
PLEASE, no work on BC other than making the cells paint its border by themself. Its waisted time. The code should be removed and not patched.
WFM with "another testcase", but still happens for "testcase". Bernd, can you do me a favor and mark this bug as depending on the bug you alluded to in comment 4?
Now "another testcase" is still WFM, and "testcase" gives me this assertion instead of the CellIterator one in the bug summary: ###!!! ASSERTION: invalid BC damage area: 'PR_FALSE', file /Users/jruderman/central/layout/tables/nsTableFrame.cpp, line 3760