Hi Alan this is indeed very strange. here the result of uname -a 4.12.14-lp150.12.28-default #1 SMP Mon Dec 3 16:46:15 UTC 2018 (b91289f) x86_64 x86_64 x86_64 GNU/Linux
My distribution is opensuse leap 15.0 and my qt is 5.12.0 (although i had exactly the same result with 5.11 as i've tested it a few weeks ago) just built again with your cmake suggestion and the results are exactly the same as before. I'l try to get an hand during next week on a computer with a different linux distribution just for testing. cheers, On Fri, Dec 14, 2018 at 11:41 PM Alan W. Irwin <alan.w.irwin1...@gmail.com> wrote: > On 2018-12-12 15:24-0000 António Rodrigues Tomé wrote: > > > Hi Alan, > > now that 5.14 is out I would like to be more useful in solving the qt5 > > driver problem. However I do not complete understand the driver system to > > be of big help.However for me I found a work arround that seems to work > > quite well. I send you 4 files, output of x01 example using qt driver, > png > > and pdf) the ones with Realease in name were created by the actual qt > > driver version 5.14 that I constructed with the command > > cmake -DCMAKE_INSTALL_PREFIX=/home/artome/plplot/PLPLOTNEW/ > > -DDEFAULT_NO_CAIRO_DEVICES=ON -DDEFAULT_NO_QT_DEVICES=OFF DENABLE_cxx=ON > > -DPL_DOUBLE=ON -DPLPLOT_USE_QT5=ON DEFAULT_NO_BINDINGS=ON ../ >& > > cmake.out > > Hi António: > > Please keep this discussion thread on list. > > I think it would be better style (and also I think -DBUILD_TEST=ON is > essential) > to use instead > cmake -DCMAKE_INSTALL_PREFIX=/home/artome/plplot/PLPLOTNEW/ > -DBUILD_TEST=ON -DDEFAULT_NO_BINDINGS=ON -DENABLE_cxx=ON -DENABLE_qt=ON > -DPLPLOT_USE_QT5=ON -DDEFAULT_NO_CAIRO_DEVICES:BOOL=ON ../ >& cmake.out > > I just tried all those options here (except for your > -DCMAKE_INSTALL_PREFIX), and they worked fine, > i.e., eliminated all bindings other than cxx and qt, dropped the cairo > devices, and included > all qt devices (which happens by default if -DENABLE_qt=ON). > > What is your exact platform (I think it might still be opensuse, but what > version?) and what are the results of the "uname -a" command on that > platform? Also, what is the version of Qt5 that you are using? > > The reason these details are important is I cannot replicate the text > alignment issues you demonstrated with pdfqt and pngqt results for > example one on your platform. My platform is Debian Buster (with the > Qt5 5.11.2 suite of libraries) installed on AMD64 hardware (Ryzen 7 > 1700). I demonstrated that I cannot replicate your badly aligned > results by configuring cmake in the way I described above and then > executing the following commands: > > # Build the prerequisites for the specific examples below > make -j16 test_c_pngqt > make -j16 test_c_pdfqt > > # Build two specific examples > examples/c/x01c -dev pngqt -o test.pngqt > examples/c/x01c -dev pdfqt -o test.pdfqt > > I have attached those two result files so you can see for yourself that > there are no text > alignment issues. > > Also your suggested changes to plqt.cpp likely conflate two issues. (i) > alignment issues caused > by your platform and not PLplot, and (ii) the size of the resulting > characters (which I > would prefer to put off discussing until later). So to check those are > two separate issues, > what happens if you restore plqt.cpp to its original form except for this > one line change: > > PLDLLIMPEXP_QT_DATA( int ) vectorize = 0; > ==> > PLDLLIMPEXP_QT_DATA( int ) vectorize = 1; > > ? I am pretty sure that is the equivalent of the first part of your > change where you > are forcing the vectorize part of the code to be executed. > > If that one-line change is all you need to bypass your Qt5 platform > issues, then I would be willing to test that simple change here to see > if it also works on my platform. But the vectorize = 1 code path has > not been tested by anybody but you over the years and is just likely a > workaround for your Qt5 platform issue so I likely won't push that > one-line change in general. > > Also you do appear to be having text alignment troubles consistently > over the years with your (opensuse?) Qt5 platforms so perhaps it is > time to try other Qt5 platforms? For example, you could build the > upstream version of Qt5 yourself. The very early versions of Qt5 did > have character alignment difficulties in that upstream version, but my > experiments showed they solved the alignment issues important to > PLplot by 5.3.x so you will likely find the latest upstream Qt5 also > does not have text alignment issues. And if that is the case, then > that experiment should be the basis of a bug report to your > distribution. Of course, if that bug report gets no response another > possibility is to try Debian (which I know works well now with Qt5) > some Debian derivative such as Ubuntu or Mint or some other rpm-based > distribution such as Fedora. > > Good luck, and let me know how it goes. > > Alan > __________________________ > Alan W. Irwin > > 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 > __________________________ -- António Rodrigues Tomé Universidade da Beira Interior Instituto D. Luís (lab associado) email address: art...@gmail.com art...@ubi.pt http://www.researcherid.com/rid/A-5681-2013
_______________________________________________ Plplot-devel mailing list Plplot-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/plplot-devel
