[Plplot-devel] RE : Alpha channel support has been added for the qt device driver

[Rochel, A. <a.r...@im...>]

Hi Alan
________________________________________
De : Alan W. Irwin [ir...@be...]
Date d'envoi : mercredi 11 mars 2009 23:36
À : Rochel, Alban
Cc : PLplot development list
Objet : Alpha channel support has been added for the qt device driver

Hi Alban:

As of revision 9726, we now have alpha channel (opacity) support for the qt
device driver.  The Qt documentation says QColor actually has 4 channels,
with the last one defaulting to completely opaque if you don't specify it
which was the mode that you originally implemented.  Fortunately, it was
easy to change from that completely opaque mode to one that takes advantage
of the opacity information supplied by PLplot.  The resulting example 30 is
much improved (e.g., shows transparency effects) for all the former
rasterqt devices (pngqt, jpgqt, etc.) and for qtwidget.  Note, we don't expect
good-looking results for page 2 of example 30 because we are trying to
approximate a gradient with a series of adjacent rectangles.  (Someday, PLplot
will have proper support for gradients, and this page will look much
nicer then.)  Spot checking of the remaining examples (all of which use
completely opaque colors) showed no issues for the former rasterqt devices
and qtwidget.

As a result of my testing of these alpha channel changes, I noticed that
epsqt, pdfqt, and svgqt (devices which were not covered by the former
rasterqt device) all had problems for filled surfaces, e.g., the solid
blocks in example 30 are not filled at all, and the surfaces in example 8
are lace-like (actually a rather cool-looking effect) rather than solid.
This issue pre-dates my transparency changes (e.g., it occurs for revision
9725). As you know, these same devices also have problems with segfaults if
-fam is not specified which indicates memory management issues for these
devices. It is difficult to tell if that is a nasty coincidence or the two
issues are related, but it is certainly possible that solution of the
memory-management problems for the 3 devices may also sort out the
surface-fill issues.

Alan
__________________________
Alan W. Irwin

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the FreeEOS equation-of-state implementation
for stellar interiors (freeeos.sf.net); PLplot scientific plotting software
package (plplot.org); the libLASi project (unifont.org/lasi); the Loads of
Linux Links project (loll.sf.net); and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

Actually, I have fixed all these alpha channel / filled surfaces problems and improved memory management in the upcoming update I should send you today. A better text management will be provided (thanks to Werner, who helped a lot), as well as various other improvements.

Cheers,

Alban