On Sat, Mar 13, 2010 at 5:12 PM, Alan W. Irwin <ir...@beluga.phys.uvic.ca> wrote: > On 2010-03-13 13:16-0800 Alan W. Irwin wrote: > >> On 2010-03-13 14:45-0500 Hazen Babcock wrote: >> >>> Alan W. Irwin wrote: >>>> >>>> Hazen, will you have a look at this issue, please? This transparent >>>> background issue does not appear for, e.g., -dev pscairo so it appears >>>> this >>>> issue may be specific within the cairo devices to just -dev xcairo >>>> although >>>> I haven't checked the other cairo device possibilities beside pscairo. >>> >>> Looks like it was an issue with the driver not using the "good PLplot >>> core >>> fallback clear capability". >> >> Hmm. after updating to get your fix and "make cairo" I still get the same >> (bad) result (with page 1 still visible underneath) for >> >> examples/c/x02c -dev xcairo -bg 0000FF_0.1 >> >> examples/c/x02c -dev qtwidget -bg 0000FF_01 >> >> was fine. >> >> I haven't tried a completely clean build yet, but "make cairo" did create >> a new cairo device driver so your fix should be properly built. [...] > > Yep, same (bad) results for an absolutely clean build tree, and I > double-checked the cairo.c code had your fix. In fact, no change in the > results is expected as a result of your change since the default value of > pls->dev_clear is zero in any case. > > Just in case, I also tried a direct (rather than X-terminal) display and > continued to get the same bad xcairo and good qtwidget transparent > background results. > > Do you currently get good results for the above test with -dev xcairo? > > Currently, there are problems with Hez's recent cairo change (see my next > post) so that the cairo file devices all segfault. However, if I revert > back to --revision 10861 for cairo.c, then the segfaults disappear and then > pngcairo does not have this transparent background issue while xcairo still > does. The conclusion from this test is the transparent background issue is > likely due to some difference in treatment between the cairo file devices > and xcairo, and whatever the xcairo issue, it is independent of Hez's recent > changes. >
I get similar results on a fresh trunk source tree. What should the default background color be? The qtwidget device seems to assume white as the default background, regardless of the base color palette: Both: ./x11c -cmap0 cmap0_default.pal -bg 0000ff_0.3 -dev qtwidget and ./x11c -cmap0 cmap0_alternate.pal -bg 0000ff_0.3 -dev qtwidget produce output with the same light-blue background color. This is probably the best approach from the standpoint of "least surprise" from the user. I have the few lines of changes ready to address this once the background color to overlay the transparent color over is decided on. Hez ------------------------------------------------------------------------------ Download Intel® Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev _______________________________________________ Plplot-devel mailing list Plplot-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/plplot-devel
