Roger Roelofs wrote: >This same question came up a week or so ago and the resolution was >the same. Put the bg image on the container centered with margin: 0 >auto; instead of on its parent. > > Yes, that's what I replied this morning in a mail to Don which by accident didn't go to the list (wrong alias...):
| Hi Don, | Because you have a fixed width and height of the header, why not use | a (combined) background-image | <http://home.tiscali.nl/developerscorner/css-discuss/images/don-bg.gif> | (not in the body, but) as bg-image in the top of the #wrap ? The he will | shrink and expand just exactly with the rest of the wrap: the white or | other colored background of the content. In the meantime, I tried a testpage <http://home.tiscali.nl/developerscorner/css-discuss/pkp-test.htm> - no more problems in IE. As far as the theoretical discussion ("why, o IE, why?") I'm guessing that the reason was: every now and then IE can have a rounding-off difference between the centering position of a bg-image in the body, and the centering position of a wrapper, when the window is made smaller. Indication: when good looking, you see not only the right side is peeking out with 1px, but at the same moment the left side is also jumping with 1px to the right. So I made a quick testpage with an image of 760px in the background (same size as the wrapper), to see if that would solve the problem. Result <http://home.tiscali.nl/developerscorner/css-discuss/pkp-test2.htm>: no! Now we see it is very systematic. Every time we move (in IE) the window 1px smaller, first the wrapper goes 1px left (bg-image stays in position and is peeking out at the right side), next step of 1px moves the bg-img 1px left (and wrapper stays in position), and so on, just in turn... francky ______________________________________________________________________ css-discuss [EMAIL PROTECTED] http://www.css-discuss.org/mailman/listinfo/css-d List wiki/FAQ -- http://css-discuss.incutio.com/ Supported by evolt.org -- http://www.evolt.org/help_support_evolt/
