On Sun, 28 Aug 2005 10:15 pm, Andreas L Delmelle wrote: > On Aug 28, 2005, at 15:57, Manuel Mall wrote: > > This is just a clarification question to those in the know. > > > > In HTML when specifying a table browsers usually choose the > > smallest width without causing unforced breaks in columns. That is > > in XSL-FO terms the ipd of the table can be smaller than the > > containing block. In the current fop version it appears as if the > > table width is always forced to the width of the containing block, > > i.e. it behaves like setting width="100%" in HTML on the table. I > > compared this to XEP and it renders more like HTML. > > > > Is this a feature, a bug, a not yet implemented, or do I > > misunderstand something? > > What layout-algorithm are you referring to exactly: fixed or auto? > I was referring to "auto" and because fop just silently did it I had the wrong impression it was implemented.
> table-layout="auto" is still unimplemented ATM. There have been > numerous loose remarks about this in various threads, but no real > work on this has started AFAIK. > > As for table-layout="fixed", CSS states that: > > "The table's width may be specified explicitly with the 'width' > property. A value of 'auto' ... means use the automatic table layout > algorithm." > http://www.w3.org/TR/REC-CSS2/tables.html#width-layout > > So, all things considered, if one doesn't specify an exact (fixed or > percentage) IPD on the table, I think this means it is currently > unimplemented. > I haven't run any tests myself, but if FOP silently supposes a > default inline-progression-dimension="100%" in this case, then that > would definitely be a bug. > > We could either: > - drop the table completely (+ warn about this, of course) > - explicitly notify the user that, because auto-layout is not > supported, the default value of "auto" is ignored and replaced by > "100%" I second that for the time being. > - throw a FOPException and exit > > > Cheers, > > Andreas Manuel