On 2013-07-05 16:08-0000 Schwartz, Steven J wrote:
> Here's the top of the (correct) linux eps:
>
> %!PS-Adobe-3.0 EPSF-3.0
> %%BoundingBox: 1 0 596 842
> %%Creator: Qt 4.8.2
> %%CreationDate: Fri Jul 5 16:40:03 2013
> %%Orientation: Landscape
> ... snip ...
> /pageinit {
> % 210 *297 mm (landscape)
> 90 rotate 1 -1 scale } def
> %%EndProlog
>
> And here's the top of the (incorrect) windows eps:
>
> %!PS-Adobe-3.0 EPSF-3.0
> %%BoundingBox: 1 0 843 595
> %%Creator: Qt 4.8.4
> %%CreationDate: Fri 5. Jul 15:25:29 2013
> %%Orientation: Landscape
> ... snip ...
> /pageinit {
> % 297 *210 mm (landscape)
> 90 rotate 1 -1 scale } def
> %%EndProlog
Hi Steve:
The above results clearly demonstrate an error in the qt device driver
bounding box results generated by PLplot on the Windows platform.
Of course, the "art" of any debugging effort is to
simplify as much as possible the test showing the bad results so
that other developers can quickly and easily reproduce the problem.
Here are parts of the PLplot code that may be relevant:
software@raven> grep -i bound \
include/qt.h drivers/qt.cpp bindings/qt_gui/plqt.cpp
bindings/qt_gui/plqt.cpp: QRectF bounding;
bindings/qt_gui/plqt.cpp: bounding = tempPainter.boundingRect( rect,
Qt::AlignHCenter | Qt::AlignVCenter | Qt::TextDontClip, text );
bindings/qt_gui/plqt.cpp: double offset = QFontMetrics(
tempPainter.font(), &tempPic ).boundingRect( text ).top(); // Distance between
the baseline and the top of the bounding box
bindings/qt_gui/plqt.cpp: path.addText( QPointF( bounding.left(),
bounding.top() - offset ), tempPainter.font(), text );
bindings/qt_gui/plqt.cpp: tempPainter.drawText( rect, Qt::AlignHCenter |
Qt::AlignVCenter | Qt::TextDontClip, text, &bounding );
bindings/qt_gui/plqt.cpp: p->drawPicture( (int) ( xOffset + bounding.width()
/ 2. ), (int) -yOffset, tempPic );
bindings/qt_gui/plqt.cpp: xOffset += bounding.width();
So assuming you want to take this further, comparisons between gdb
runs on the Linux version and Windows version might show you that,
e.g., tempPainter.boundingRect was returning a different result on the
two systems. Which could lead to a simple test case containing a call
to Qpainter.boundingRect that you could approach the Qt developers
with.
If Qpainter.boundingRect is the issue, full documentation of that
function occurs at
<http://harmattan-dev.nokia.com/docs/library/html/qt4/qpainter.htm>.
HTH.
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); the Time
Ephemerides project (timeephem.sf.net); PLplot scientific plotting
software package (plplot.sf.net); 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
__________________________
------------------------------------------------------------------------------
This SF.net email is sponsored by Windows:
Build for Windows Store.
http://p.sf.net/sfu/windows-dev2dev
_______________________________________________
Plplot-devel mailing list
Plplot-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/plplot-devel
