Hi Phil: This is mostly addressed to you because it turns out one of your commits is causing release-critical trouble with the plsori call in example 3 as well as any attempt to use a non-zero -ori option.
The current status of the PLplot-5.14.0 release process is as follows: The PLplot development freeze that started on October 27th contines, i.e., do not push any topics other than documentation updates or well-tested bug fixes. I had gotten to the stage in README.Release_Manager_Cookbook where I was checking a locally generated version of our website when I discovered all lines were missing from example 3 for pngcairo. Further checking revealed that if plsori is called with a non-zero value (example 3) or the -ori option is used with a non-zero value (all other examples), then all graphical elements in the plot disappear for all device drivers! The plsori routine is simply a wrapper for plsdiori, and the -ori command-line option calls plsdiori directly. So this changed behaviour for plsdiori since 5.13.0 is a substantial regression which is therefore release critical. I used the commands make -j16 x01c ; make -j16 cairo examples/c/x01c -dev xcairo -ori 0.001 to help git bisect the issue, and from that bisection the first commit to cause missing graphics (i.e., missing lines) for that last command (or the same command for any other device) was your commit plplot-5.13.0-47-g124a0c3a2, with the subject line: "Set all clip limit changes I could find to use plP_sclp and hence get recorded in the buffer" Furthermore, for a local topic branch based on master tip, if I locally reverse that commit, the problem with missing graphics for the above example is solved and similarly (note that example 3 contains a call to plsori) examples/c/x03c -dev xcairo (or any other device) works fine (without missing graphics) as well. With the reversed commit, one thing that does not work well is (as expected) examples/c/x33c -dev wxwidgets which has rendering issues (missing lines and other graphical items). However, my test of that example with -dev xcairo, -dev qtwidget, and -dev xwin do render well with this reversed commit. @Phil: This commit (or its reverse) is not intrusive at all so it should not be that difficult to figure out why this change had such a catastrophic effect on plsdiori. N.B. Once I have that figured out, I will likely just go ahead and reverse your commit so that -ori and calls to plsori continue to work properly for our users just like they did for release 5.13.0. Of course, reversing your commit means that your attempt to record clip limits in the buffer this way in order to fix "examples/c/x33c -dev wxwidgets" will have to be addressed post-release in a way that preserves the functionality of -ori and plsori, but likely that will be far more intrusive than your commit plplot-5.13.0-47-g124a0c3a2 so it very likely should be a post-release change in any case. In sum, if I don't hear back from you quickly, I will likely just go ahead and reverse your commit tomorrow (Tuesday) and comprehensively test that change in anticipation of making the release later tomorrow. More later. 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 __________________________ _______________________________________________ Plplot-devel mailing list Plplot-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/plplot-devel
