Jean-Marc Lasgouttes wrote:
"Martin" == Martin Vermeer <[EMAIL PROTECTED]> writes:
Martin> Jean-Marc, should this go into branch too? - Martin
Well, it does not fit with the definition of Wide() in 1.4.x.
Conerning the need for the fix, I think it just hides the bug in a
particular case. The real fix would be to manage to get rid of this
wide() thing :)
Indeed. AFAIU, the wide() property was introduced by Martin in order to
be able to use the singlePar optimisation within insets. This was
required because the rowpainter does not distinguish between text
drawing and decoration drawing.
So the correct fix is to cleanup the rowpainter. I won't do this for 1.5
but, if I am still active for 1.6, I will do it.
It is possible to set paragraphs centered in almost all text insets,
so blocking it in floats seems a bit vain. The bug remains for
branches for example, and I do not think it would be a good idea to
disable wide() in branches just for that.
What is the correct behavior we would like to obtain for centered
paragraphs in a wide inset, BTW? Would it be really difficult to
change metrics computation so that a centered paragraph always tries
to occupy the whole screen width (like with a hfill)? I think this
would solve the problem in a better way.
I don't think so. I don't wanted all my text to be completely
re-organized when I enter the inset.
Abdel.
