On 2015-03-09 12:22-0000 Phil Rosenberg wrote: > Hi Alan > > Here is a status update on wxWidgets, but the headline is that I now > feel that the wxWidgets driver is at an appropriate stage for release.
That is excellent news indeed! I still have lots on my PLplot agenda so the headline from my perspective is the release date is still not settled, but I hope to firm up the date late this week. There are, of course, still a lot of wxwidgets "would be nices" that we jointly mention below, but nothing release critical as far as I am concerned because of the USE_OLD_WXWIDGETS cmake option I discussed with you (not implemented yet, but I plan to). For the others here this option will give the users the chance to choose the old wxwidgets device, associated library, etc., at cmake time (not run time where there would be a lot of clash issues to sort out between old and new wxwidgets device). So if one of the issues below (or some other issue we are unaware of) is a showstopper for them, we will have provided them a CMake option to go back to using the old (deprecated) wxwidgets code for at least this release. More below in context. > > The issues that still exist are on trello, so you can view them is you > want. However I now think that the new driver is of generally the same > quality as the previous one. > > Of the remaining issues the ones that I feel are of highest priority are > > 1) The preservation of the bop state in the buffer > 2) 3d text is not quite correctly aligned. > 3) Adding wxStream constuctors that mirror plstream constructors. > 4) There are some aspect ratio issues with x03, but only when resized > and only on Linux. I think this is related to the rotated orientation. > > My understanding is that Jim is doing 1), but if he is busy with other > stuff I'd be happy to sort that. Please take a look at his patch series he sent this weekend. One of those patches apparently was to fix a wxwidgets issue so part/all of that series may be exactly what you need, but I am leaving it to you to evaluate which of his commits should be applied to master now, and which should wait until post-release (if they don't affect wxwidgets). 3) is straightforward and I will do > that this evening. 2) is a bit trickier as I don't really have a good > understanding of why it is wrong - it uses the same maths as the old > driver, but I wonder if the aspect ratio code has changed (for fixed > aspect) and there is some similar change required in the text. As > issue 4 is a rotated plot I consider this a bit of an edge case for an > interactive driver - it does need fixing, but I can live with it in > the short term. > > If you can live with 2 and 4, then we are pretty well ready to go. If you can live with them for now, then I can too. However, I would appreciate you figuring out what was wrong with my patch to fix the wide legend box issues for the unicode case. At the time, you said you were getting complaints from Windows about unicode fonts, but since my patch simply used a facility you had already developed to determine physical width and height of the string that would be rendered by the wxwidgets libraries, I suspect that issue was attributable to other Windows issues you were having at that exact same time so please take another look. > One question I had though, is there any different behaviour required > for the no pause state? Currently the driver does not pause at all > regardless of this state. It sends all its data to the viewer then > continues - for the examples it then exits as it reaches the end of > main(), but leaves the viewer open with the plots. I think the fundamental issue here is that new wxwidgets launches a viewer that runs in a completely detached state. That can be good in certain circumstances, but it obviously goes against the paradigm that is used for every other interactive device. So I believe the solution here is to implement a wxwidgets driver option that allows detached mode (as now), but by default operates in attached mode. Once attached mode works, then that makes it possible to control the plot directly from the example, e.g., by specifying the nopause option. I assume attached mode would also allow -locate mode to work for example 1, sort out some of the example 17 issues, etc. In sum, all of the above are "would-be-nices" in my opinion rather than showstoppers. So you are free to tackle as few or many of them as you like in the next few days, and we should look again where we jointly stand near the end of this week. 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 __________________________ ------------------------------------------------------------------------------ Dive into the World of Parallel Programming The Go Parallel Website, sponsored by Intel and developed in partnership with Slashdot Media, is your hub for all things parallel software development, from weekly thought leadership blogs to news, videos, case studies, tutorials and more. Take a look and join the conversation now. http://goparallel.sourceforge.net/ _______________________________________________ Plplot-devel mailing list Plplot-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/plplot-devel
