Abdelrazak Younes wrote: > Alfredo Braunstein wrote: >> Abdelrazak Younes wrote: >> >>> Alfredo Braunstein wrote: >>>> Abdelrazak Younes wrote: >>>> >>>>> I just now understand what you mean... But it is too late now to cure >>>>> that, maybe tomorrow... >>>> No worries :-) >>> It was pretty easy to solve actually, the moment you know how to solve >>> it I mean... >> >> I still think that it's more consistent to always pass around baseline y >> coordinates, even if this implies substracting/adding ascent a couple of >> times (0 performance impact I suppose). > > There's as small performance impact because we do less lookup of the > relevant ParagraphMetrics. And it's less lines of code too. > > >> But I guess this is matter of >> preferences... > > I still disagree ;-) > My rationale is that the default should be: "draw in this frame". The > fact that the inner text should be aligned with the outer text is not > the problem of the inner text. The containing inset should make sure > that the inner text is properly aligned. That's the main reason for the > change.
There's no conceptual difference IMO. You say that the parent inset says to the child: "draw in this frame", but this is because he knows where the baseline of the inset is and he computes the frame accordingly. The same thing could be done on the other way around: "draw in this baseline", because he knows anyways in which frame he will draw. The fact is that the outer inset knows everything so either passing the baseline or the frame conveys all the information. I'm not religious about this, but now we have some methods that receive baselines and some others that receive frames... A/
