Re: [Plplot-general] Plplot on Mingw

[Walt B. <wal...@gm...>]

SUCCESS!

The critical environment variable is PLPLOT_LIB, not that ENV thingie.

I found it in the excellent documentation. I know--that is cheating,
but I should have tried that again before pestering you guys.

Another question: if I (or you) provide this install directory as a
Windows binary distribution, how many systems do you think it
would work on?

Let's hope this continues to work :=). And thanks again for the help.

On Fri, Nov 28, 2014 at 3:26 AM, Alan W. Irwin <ir...@be...>
wrote:

> On 2014-11-28 08:30-0000 Arjen Markus wrote:
>
>  Hi Walt,
>>
>>
>>
>> I have encountered such messages myself - they are due to the PLplot
>> start up routines not being able to find the dynamic drivers (indeed they
>> reside in the DLLs in the "drivers" directory). Within the build tree, it
>> is easy enough for the routines to find them, but outside you may need to
>> set the environment variable PLPLOT_DRV_DIR.
>>
>>
>>
>> A further problem you would encounter is the location of the fonts and
>> some other files. For this set the environment variable PLPLOT_LIB_ENV to
>> the directory containing the .fnt and .al files. (Note that default values
>> are used that correspond to the installation prefix, but if there is some
>> discrepancy, then these environment variables will help out).
>>
>>
>>
>> A small caveat: I have not actually tried all of this just now (I scanned
>> the source code for the correct names), but in the not-so-distant past I
>> had to resort to this method, so I am confident it works. If not, let me
>> know and I will dive into it. Having done it before means I should be able
>> to find it quickly.
>>
>
> Hi Arjen and Walt:
>
> @Arjen: Just to clarify, I haven't tried extensive testing of the
> "MinGW Makefiles" generator case, and I know there are continuing
> PLplot problems with blanks in pathnames.  So for either of those
> cases, the above historical workarounds (or some other workaround)
> might be needed for Windows.  And also for 5.10.0, the "MSYS
> Makefiles" case with no blanks in pathnames passed some limited build
> tree tests but not install-tree tests.
>
> However, I just double-checked and for the git master branch (at least
> for commit id 3e817e8 on that branch that came after the release of
> 5.10.0,see details at the bottom of
> <https://sourceforge.net/p/plplot/wiki/Testing_PLplot/>), I finally
> got all comprehensive testing to work without issues for the the "MSYS
> Makefiles" case with no blanks in pathnames case.
>
> @Walt: I rarely do comprehensive testing on my Wine platform because
> it is an incredibly slow version of Windows (comprehensive tests take
> a couple of days of cpu time rather than a couple of hours). However,
> I don't anticipate that we have introduced any Windows build or test
> issue on the master branch since the last time I did that
> comprehensive test. Therefore, I suggest you use the latest git master
> branch version of PLplot and the "MinGW MSYS" generator with no blanks
> in any pathname to see if you can replicate the good Windows result I
> had for that case for an earlier master branch version.
>
>
> 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
> __________________________
>



-- 
Walt Brainerd