[Plplot-devel] epa_build now gives a clean comprehensive test result on MinGW/MSYS for "plplot-lite"

[Alan W. I. <ir...@be...>]

Hi Arjen:

As of revision 13051, I have implemented a typographical error fix
that allows the test_tclsh_standard_examples target to work on
MinGW/MSYS.  When this change is combined with a previous commit
(which for MinGW/MSYS continued to enable the
test_pltcl_standard_examples target, but which removed that target
from the list of overall interactive test dependencies because of
<http://sourceforge.net/p/plplot/bugs/144/>) the result is an
epa_build of the build_plplot_lite target using
-DCOMPREHENSIVE_PLPLOT_TEST_INTERACTIVE=ON works for the very first
time on my MinGW/MSYS/Wine platform.  When this is combined with my
previous (just before the release of plplot-5.10.0) good results for
an epa_build of the build_plplot_lite target using
-DCOMPREHENSIVE_PLPLOT_TEST_NONINTERACTIVE=ON, it appears we now have
a clean bill of health for the "lite" version (excluding qt, cairo,
wxwidgets, octave, and X) of PLplot on MinGW/MSYS.

To do such a complete test on Wine takes me several days of cpu time
because of the horrible Wine startup latency issues I have remarked on
in the past.  So I prefer not to do such a test very often when the
alternative for someone else with access to the Microsoft version of
Windows should be an hour or so of cpu time.  (On Linux, the
equivalent tests of a complete version of PLplot take roughly an hour
of cpu time so I assume the Microsoft Windows platform will finish
comprehensive tests of the lite version of PLplot in substantially
less than an hour.)

Therefore, will you please confirm these results on your Microsoft
version of Windows, and after that please confirm
http://sourceforge.net/p/plplot/bugs/144/ on that platform to make
sure that bug is not due to some issue with my Wine platform?

We are still early in the release cycle for the forthcoming release of
PLplot-5.11.0 so there may be a tendency on your part to assume
testing is not important at this stage.  But on the contrary I think
it is most important now to establish some definitive test benchmarks
for the MinGW/MSYS/Microsoft Windows platform. If you are willing to
do that, then near the end of this release cycle you will only have to
repeat these "canned" tests for substantially less thanan hour of cpu
time to assure that no regressions have been introduced for the
MinGW/MSYS platform by any changes we make in this release cycle. That
is a much better situation than me performing these tests on Wine at
the end of this release cycle which would substantially delay the
release because of the days of cpu time that would require.

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
__________________________