Renzo, How about attaching it to http://issues.apache.org/jira/browse/ADFFACES-132
That bug covers the Safari problems, but this strategy may work well for Safari too, though playing around with: http://www.quirksmode.org/css/display.html ... suggests that we might need to use display: table-row somewhere as well. -- Adam On 1/25/07, Renzo Tomaselli <[EMAIL PROTECTED]> wrote:
Finally I got a working solution: - the table representing a single tab must have style="display: table-cell; border-spacing:0" instead of style="display: inline". - its tbody does not need any attribute. - these modifications concern only Gecko. This way tabs are displayed properly on Firefox 2.0. It would be interesting to try out other browsers, I only have IE7 and FF 2.0. I can send a patched NavigationPaneRenderer.java to anyone interested. -- Renzo Renzo Tomaselli wrote: > Adam, indeed it doesn't work with multiple tabs, since the second one > is rendered just below the first one (out of div). > Code comments state that inline on table was needed for both IE and > Firefox 1.5, while inline on tbody is needed on Firefox 1.5. > I will investigate further, the only other chance I'm left with is to > have a dummy nop extra-tab, without icon. This way rendering is ok on > Firefox, but the overall stuff is so much unstable that I don't feel > much safe. > > -- Renzo > > Adam Winer wrote: >> I believe the display:inline is necessary for at least one >> browser (IE, I think?), so removing it isn't OK. navigationPane >> is definitely in need of an HTML overhaul, though - it >> renders pretty badly on Safari. >> >> -- Adam >> >> >> On 1/24/07, Renzo Tomaselli <[EMAIL PROTECTED]> wrote: >> >>> After a long stripping session, I came up with a simple html page, >>> derived from a rendered NavigationPanel tab. >>> After placing this page on the local filesystem together with a >>> stripped >>> css, Common11-m7.js and all involved icons (tab3-*), Firefox 2.0 >>> displays this page completely only the first time or after any >>> reloading. >>> It does not complete the page after clicking on the link. No console >>> errors, though. >>> It seems pausing in the middle. Rendering completes if clicking either >>> buttons. Even activating plugins such as CSSviewer or Firebug completes >>> the page. >>> Please notice that I removed all inline script function calls, so that >>> no Trinidad js functions should be called. However, if I remove the >>> script loading, page shows up ok. If I remove the outermost table, it's >>> ok. If I place all those files on a remote server then I browse that >>> page from here, it's ok. >>> If I remove the onclick attribute from the link, it's ok (it's just >>> submitting the enclosing form, the original page invoked another >>> function). >>> >>> Now the funny point: look at those initial table/tbody elements having >>> "display: inline;". If I remove *both* attributes, rendering is ok ! >>> Don't ask me why. >>> Now I can easily patch NavigationPaneRenderer again (third time in a >>> week) to remove those attributes, but does anybody know about any >>> potential side effects ? >>> Besides this page, I will check the real application ASAP through a >>> patched NavigationPaneRenderer. >>> In any case, here are all data (js and icons are from Trinidad): >>> > >
