This is a meta bug that groups all bugs that deal with incorrect heights on tables, row groups, rows, cells.

In the attachment, IE honors the row with height=400 when the table has height=200, but does not honor the 2nd row with height=80%. It seems to place the constraint that the sum of the percentage rows cannot exceed 100%, but places no constraint on pixel valued rows. Is this reasonable, or should the 2nd row with height=80% get a full 80% of the table height? If possible, I would like to avoid the complicated, ess-efficient balancing that goes on with columns where the table width constraint has the highest priority. It appears that IE has avoided it as well.

I think you need to initialize the new mBits in the root reflow state constructor, no? Also, mBits seems like a pretty ambiguous name, why not something more meaningful like mRefowFlags? Also, please indicate what testing you have done, which cases it fixes and which it does not fix. Thanks.

reassigning to m0.9.6

Need to change unsigned to PRUint16 in struct with bitfields (from alexsavulov, attinasi).

Comment on attachment 56516 [details] [diff] [review] more up to date patch with reviewer's suggestions sr=attinasi, r=alexsavulov

checked in

This checkin caused bug 109043