On Fri, Dec 09, 2005 at 11:21:29AM +0100, Georg Baum wrote: > Jean-Marc Lasgouttes wrote: > > >>>>>> "Georg" == Georg Baum > >>>>>> <[EMAIL PROTECTED]> > >>>>>> writes: > > > > Georg> I wanted to be on the safe side and change as little as > > Georg> possible. Of course I could also iterate over the cells of each > > Georg> inset containing text, but I am not sure if this is the right > > Georg> thing to do (e.g. for collapsable insets or ERT). Or I could > > Georg> move the code of InsetTabular::setFont into the loop in > > Georg> setCharFont (or setInsetFont if that is used). What do you > > Georg> think? > > > > I would tend to think that all collapsable insets should see their > > content changed too. The case of ERT is a bit problematic, but we > > could maybe try to detect it. > > I investigated a bit, and have now a solution. Since we have two problems > here, I'd like to fix them one after one. > > The first problem which has not beed addressed yet is that insets that > return true in noFontChange() inherit the environment font on screen but > don't inherit it in the .tex output. This is fixed by the attached patch. > OK to apply? > > The second problem is that font changes need to be propagated to these > insets. I'd like to address this problem later. > It works as advertised for my testcase. Now the screen and the output is in sync all the time. Applying a change over a selection that includes a table in the middle won't affect the table cells, but at least that fact is plainly visible on screen. This is an improvement, there is no confusion any more.
Helge Hafting
