Description of problem: Icons aren't positioned well over spreadsheet cells. Version-Release number of selected component (if applicable): openoffice.org-core-2.3.0-6.11.fc8 How reproducible: Easy. Steps to Reproduce: 1. Load attached test case in oocalc Actual results: The "traffic lights" icons in column C are increasingly offset by a small amount with each table row so that they're off by about one row in rows 200 and higher. Expected results: The icons are positioned well within each cell. Additional info: The document file was exported from an SAP system (i.e. not created by OOo).
Created attachment 293248 [details] ODS document exhibiting the problem
Created attachment 293249 [details] screenshot of the top cells with icons (small offset visible already)
Created attachment 293251 [details] screenshot of the bottom cells with icons (offset is about a row here)
I think the substantive problem is due to the hyperlinks affecting the height of the row, presumably as a guess for now with the positioning of the graphics are being calculated without the extra hyperlink border or something being factored in. If you made a run without hyperlinks in the output you'll probably get correct positioning. There are a few nasty things in the sample file which I've stripped out of my testing set that might still be an issue that you should probably investigate in the output engine of your xml e.g. a big blob of junk-looking text beginning with MYSTRING etc, graphics with 0 width and height. And no pretty printed xml (as you get from OOos tools->options->save disable size optimization) makes things hard as well to work with.
Created attachment 293340 [details] spreadsheet where graphics are positioned nicely
Created attachment 293341 [details] same spreadsheet where the links are formal hyperlinks and where we simply sed in some "a href"s around the http://www.web.de and otherwise untouched and the graphics get misaligned
oh and each graphic has the same z-level as every one other, they should be different so that there is a stacking order of graphics so that if they *do* overlap that one can go above or under the other.
Hmm, and can we confirm that all is well when viewed with file->page preview ?
Hmm, I don't see a quick fix for this. There are some very whacky to my mind calculations in calc for determining the height of a non-editengine cell height. i.e. all the other ones *without* a hyperlink in them. A proto patch logged on the upstream bug, and with it in place the layout is fine to the user eye. But need some feedback from calc guys as to what their intent is to see which needs to be fixed, normal cell height, or edit-engine cell height.