On Tue, 18 Jun 2002, Tino Schwarze wrote:
> On Tue, Jun 18, 2002 at 11:09:04PM +0200, Cyrille Chepelov wrote:
>
>> > Well, just let's not store the bounding box at all but let the UML
>> > Class figure it out with the help of the renderer (it needs the
>> > renderer anyway). IMHO this would require nothing more than an
>> > invisible "redraw all" cycle after loading the file.
>>
>> Sure. and of course, you never have any Connection (OrthConn, likely)
>> linked to a connection point on the right side of the UML Class objects.
>
> If the bounding box is "wrong" in the file, the connection gets messed
> up anyway the first time someone clicks on the UML Cass.
I agree that the BB shouldn't be stored in the XML file. There's simply no
use for it there.
>> No matter how we look at the problem, we need to come up with a way of
>> having a stable, reliable and repeatable way of computing text extents,
>> which should never depend on the scaling factor.
>
> Agreed. But really: I don't think there is a reliable way of computing
> text extends without depending on the actual resolution. There is a lot
> of effort in fonts to accomodate different resolutions. And as far as I
> know, fonts are always bound to specific resolutions and can change
> their extents quite unexpectedly when switching to the next
> bigger/smaller font size. Hopefully I'm a bit pessimistic here...
I would hope there's a way to turn that stuff off. Sure, we'll get less
pretty fonts, but that doesn't matter as long as they are consistent.
> Maybe we could simply ask the Pango people about that? They should be
> very competent in font issues.
They should. That would be the place to go.
-Lars
--
Lars Clausen (http://shasta.cs.uiuc.edu/~lrclause)| Hårdgrim of Numenor
"I do not agree with a word that you say, but I |----------------------------
will defend to the death your right to say it." | Where are we going, and
--Evelyn Beatrice Hall paraphrasing Voltaire | what's with the handbasket?