Occurs on M7 release, under NT 4 Workstation. At the top of the Bluemountain page, there is a blue navigation bar consisting of two tabs. Underneath this should be a thin yellow line. On M7 this line does not appear. (Ignore the blue line that appears instead -- this is a manifestation of bug 5900.) Will attach test case to simplify the problem. The test case only shows the problem when both: * the table cell contains the "nowrap" attribute * the spacer image inside the cell is right aligned (floated) This code behaves correctly on IE5 and Opera 3.6

Created a new test case, this time relating to the use of spacer elements within an HTML table. The code snippet is modified from bug 3312. The test consists of three tables. Any browser that supports both table height and spacer elements should render the three tables identically. The first table shows the background colour appearing for the entire second row, but for neither the first nor third rows. Both first and third rows have been drawn with the correct cell height. The third row has not drawn its border separating it from the second row. The second table shows the table normally, with all cells having their background painted correctly. The third table paints the background colour on the first cell of the first and third rows, and for the entire second row. The only difference between the rows here is the text that appears in the second row. The problem appears to be that the code handling the painting of cell backgrounds can't tell correctly when the cells in question have been collapsed. Changing the summary field to reflect the new test case.

[TESTCASE] The two attachements appear to be adequate test cases.

The first testcase now works ok. See also bug 10036 for similar issues with other elements, such as <br>.

Moving to M13.

mass move to m14.

Bulk moving [testcase] code to new testcase keyword. Sorry for the spam!

The url and attachements look ok compared to IE. The 2nd attachement does give a different width to the cells with images which can't be found. ChrisD if you think this should keep the bug open, please reassign to Kipp's bug list.

5.0 is rendering the 3 tables identically in 2/9 build. Verifying WORKSFORME.