In the current build of Mozilla several consecutive & nbsp; do not split at a line break, in the case of this URL, in a table cell, causing the cell to stretch to fit, messing up the other column widths to the left of the cell as well as to the right in the table. Currently Navigator pushes columns to the right of that column farther to the right and widens the table. IE5 shrinks the columns to the right to fit the table width (as defined in <TABLE WIDTH=###>

The issue has nothing to do with &nbsp's - those provide the space betwen the green column and the main table area where most of the text content is. The bug is in the green column - its too wide.

Moving to M6.

This is the same situation as bug #5835. You could mark it as a dup or keep it for now for the test case (slightly more involved than bug #5835). Attachment to follow.

Moving to M8

Marked with [TESTCASE].

*** Bug 5835 has been marked as a duplicate of this bug. ***

Here is the testcase from bug #5835 (URL was http://www.book.ru/snk/) http://bugzilla.mozilla.org/showattachment.cgi?attach_id=109 I should write it again, as it is not really clear on first reading.

Sounds very simular to bug#5797

*** Bug 1095 has been marked as a duplicate of this bug. ***

*** Bug 2947 has been marked as a duplicate of this bug. ***

[TESTCASE] It is not only an IMG in a colspan that distorts the width -- a colspan with a width will cause an identical problem. It looks like images in cells are having a direct effect on the cell's width, even if the cell should not contribute anything towards the distribution of width in a table. Thus when you set the width of a colspanned cell using width instead of a table, the same problem appears. Some more test cases, these produced by distilling bug 1095 and bug 2947: http://www.bath.ac.uk/%7Epy8ieh/internet/projects/mozilla/tablewidth.html http://www.bath.ac.uk/%7Epy8ieh/internet/projects/mozilla/tablesizing.html Note that in some cases, it is the _table_ width that is distorted rather than the simply the column width distribution.

From bug 2947, http://www.webstandards.org/ is also a relevant test case. I think there were other bugs on the same problem with that page, which was fixed and then reappeared.

I would say that http://www.bath.ac.uk/%7Epy8ieh/internet/projects/mozilla/tablewidth.html is the canonical test for colspans and width distribution. I've noted a few other extant bugs with {colspan-width} in the summary.

*** Bug 2954 has been marked as a duplicate of this bug. ***

Moving to M9.

*** Bug 10039 has been marked as a duplicate of this bug. ***

The url and attachment look like Nav4.6 (except that the images in the attachment are not found and we handle that differently). A lot has happended since I looked at this last. Marking worksforme.

Works for me too. Marking as verified works for me in the Sept 16th build.