On Sunday 08 Dec 2013 12:44:18 Alan W. Irwin wrote: > On 2013-12-07 20:14-0500 Hezekiah M. Carty wrote: > > On Fri, Nov 29, 2013 at 10:36 AM, Andrew Ross > > > > <andrewr...@users.sourceforge.net> wrote: > >> On Thu, Nov 28, 2013 at 10:56:10PM -0800, Alan W. Irwin wrote: > >>> ocaml > >>> > >>> Missing examples : > >>> Differing postscript output : 16 22 33 > >>> Missing stdout : > >>> > >>> Differing stdout : > >>> I expect those differences will be substantially reduced or even > >>> eliminated in the next few days if everyone cooperates for their > >>> favorite language(s) that still have remaining propagation issues. > >> > >> The remaining languages need updating. I will look at tcl next. I am > >> assuming Hez will take on the ocaml changes along with the changes to > >> examples 16 and 33. If someone is able to look at ada that would be > >> great, otherwise I'll try and get round to it, but my ada knowledge is > >> pretty hazy. > > > > I updated the OCaml bindings and example 22 to match the C version. I > > am hoping to have time early this week to finish the rest of the OCaml > > example updates. If not, the OCaml examples will just have to ship > > out of sync with the C versions. > > Thanks for that information which gives me a better idea of what to > expect as release manager. Good luck with your attempt to finish > off the remaining OCaml propagation issues early this week!
Alan, Having finished the changes to example 22 I proposed to do for the release I've done some thorough testing today. Firstly I compiled with a complete set of gcc flags to set the language standards for C/C++/Fortran and to enable various warnings. This showed up a number of issues which I've addressed. There are still _lots_ of warnings in swig generated code. I've tried to deal with the ones due to our swig wrappers, but most are in generic swig code. There are also a fair amount due to interaction with other libraries (wxwidgets / tcl seem particularly noisy). There remains a small set of warnings in our own plplot code. These are mostly due to design decisions made a long time ago regarding the use of PLPointer etc. By casting pointers to void * and back we lose any type and attribute checking and gcc (rightly) issues warnings. I've updated the recommended posix standards version to 2008 to deal with a couple of functions not in the 2001 posix standard. One set of warnings I've not noticed before is from gfortran. By default it warns about testing equality / inequality between real numbers. It is true this is potential liable to rounding errors causing unexpected results, but quite a lot of our examples do this (for all languages) and they produce consistent results across the different languages. I've also done a full build and run test_interactive and test_noninteractive in the build tree for both current Ubuntu and Debian unstable. Both compiled and ran without any obvious issues. I'll do some further testing, but unless I find any issues I don't propose doing any further work before the release. Andrew ------------------------------------------------------------------------------ Sponsored by Intel(R) XDK Develop, test and display web and hybrid apps with a single code base. Download it for free now! http://pubads.g.doubleclick.net/gampad/clk?id=111408631&iu=/4140/ostg.clktrk _______________________________________________ Plplot-devel mailing list Plplot-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/plplot-devel
