Le jeudi 19 juin 2008 à 10:29 +1000, Martin Sevior a écrit :
> On Thu, Jun 19, 2008 at 4:50 AM, Tomas Frydrych <tf@o-hand.com> wrote:
> > Martin Sevior wrote:
> >> Hi Everyone,
> >> I've just committed a fix to STABLE to fix this.
> >> Still needs some fixes to make equations print correctly but otherwise
> >> it is all good. We get the same layout no matter how start the initial
> >> zoom and we get consistent printing with and without X11.
> >>
> >> This code essentially mirrors what did prior to 2.6 by making all
> >> layout calculations with a fonts calculated at resolutions of 1440
> >> dpi.
> >
> > It might be worth to check the impact of this change on memory
> > footprint; iirc, PangoGlyphString takes up around 24 bytes per glyph.
> > Also, I think with the upcoming Cairo graphics class, it would be worth
> > to investigate whether zooming could be done via transforms rather than
> > via the font-size hack.
> >
>
> Hi Tomas,
> Yes this will use more more memory. At least 24 bytes
> per glyph as you say. The benchmark as to whether Rob succeeds will be
> whether he can avoid bugs 11664 and 11668 without the 1440 dpi
> reference font.
>
> Cheers
>
> Martin
>
> > Tomas
> >
I experimented on cairo zooming and text last year (for gchempaint where
I need an exact position for all glyphs). Using cairo transforms do not
work as I hoped it should do. The string sizes are not the same at all
zoom levels. I'm suspecting that pangocairo changes the font size
according to the zoom level. At this moment, I have no solution, but
plan to use coords for each glyph and draw them one by one (gchempaint
is not a word processor, so it might not be a serious performance
issue).
Cheers,
Jean
Cheers,
Jean
Received on Thu Jun 19 09:48:52 2008
This archive was generated by hypermail 2.1.8 : Thu Jun 19 2008 - 09:48:52 CEST