[Plplot-devel] MinGW/MSYS PLplot test results on Wine

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

Hi Arjen:

I need PLplot to be installed on Wine so I can test another project
(te_gen) that requires a lot of PostScript plotting on that platform.
So today I tried building and testing PLplot svn trunk version on
MinGW/MSYS/Wine (for Wine-1.5.15 and for MinGW/MSYS you can obtain
using the 20110802 installer without updates which corresponds to gcc
4.5.2).

My Wine platform did have access to version 4.5.2 of gcc, g++, and gfortran via MinGW
and a bootstrapped build of CMake2.8.9 (done with MinGW/MSYS) and
downloaded binary versions of swig and Python (a slightly dated
version 2.7.1 which I downloaded last year from
http://www.python.org/ftp/python/2.7.1), but virtually nothing else.
A default build (no -DENABLE_* or -DPLD_* options) went into an
infinite loop in the FindLua.cmake module.

Is that infinite loop a Wine peculiarity or can you confirm that same
issue there on Windows using your usual CMake command, but with
-DENABLE_lua=ON?  I am presuming here that you actually don't have Lua
installed which is also the case for my current Wine platform.  That
case presumably exercises CMake logic that ends up as an infinite loop
here.

To avoid the infinite loop I specified -DENABLE_lua=OFF, and the cmake
command (done with -DBUILD_TEST=ON) finished with no issues (although
with most components/devices disabled).

Here is the summary:

ENABLE_DYNDRIVERS:      ON
DRIVERS_LIST: mem;null;ps;svg;wingcc;xfig
DEVICES_LIST: mem;null;ps;svg;wingcc;xfig

Library options:
BUILD_SHARED_LIBS:      ON              PL_DOUBLE:      ON

Optional libraries:
PL_HAVE_QHULL:          OFF             WITH_CSA:       ON
HAVE_FREETYPE:                          PL_HAVE_PTHREAD: 
HAVE_AGG:                               HAVE_SHAPELIB:  OFF

Language Bindings:
ENABLE_f77:             OFF             ENABLE_f95:             ON
ENABLE_cxx:             ON              ENABLE_java:            OFF
ENABLE_python:          ON              ENABLE_octave:          OFF
ENABLE_tcl:             OFF             ENABLE_itcl:            OFF
ENABLE_tk:              OFF             ENABLE_itk:             OFF
ENABLE_pdl:             OFF             ENABLE_wxwidgets:       OFF
ENABLE_ada:             OFF             ENABLE_d:               OFF
ENABLE_ocaml:           OFF             ENABLE_lua:             OFF
ENABLE_qt:              OFF             ENABLE_pyqt4:           OFF

I then went on to do a test with the test_noninteractive target
and the Fortran 95 examples bombed out with the following error
message:

[ 83%] [ 83%] Fatal Error: Reading module plplotp at line 1097 column
3: Expected string
gfortran.exe: Internal error: Aborted (program f951)

I believe this is a MinGW on Wine peculiarity (I have run into this
Fortran module parsing issue before for MinGW on Wine), but could you
please run the test_noninteractive target there for the -G "MSYS
Makefiles" case to confirm the problem does not exist on your
Microsoft platform?  And when you report those results will you please
include the gfortran version string?  Here under MSYS bash.exe on the Wine
platform that version string is

bash.exe-3.1$ gfortran --version
GNU Fortran (GCC) 4.5.2
Copyright (C) 2010 Free Software Foundation, Inc.
....

A subsequent fresh build and test (with the test_noninteractive
target) using the CMake options

-DDEFAULT_NO_BINDINGS=ON -DENABLE_cxx=ON -DENABLE_python=ON
-DENABLE_f77=ON

worked without obvious run-time issues for me on Wine.  However, the diff results
from that test were

c++
   Missing examples            :
   Differing postscript output :
   Missing stdout              :
   Differing stdout            : 
f77
   Missing examples            :  00
   Differing postscript output :  14a 21
   Missing stdout              :
   Differing stdout            : 
python
   Missing examples            :
   Differing postscript output :  00 03 04 05 06 08 09 11 12 14a 15 16 17 18 19 20 21 22 23 25 26 27 29 33
   Missing stdout              :
   Differing stdout            :  23

I don't remember whether I have ever tested Python on Wine before.
Does that bad result jibe with your experience?  More later on those
Python on Windows (or Wine?) issues once I look more thorougly at them.

The few f77 issues are likely because you haven't been maintaining
those bindings/examples. I think you should continue to ignore
maintaining f77 since it is deprecated.  However, Wine is a platform
(because of the above problem with parsing Fortran 95 module files)
where only the Fortran 77 version of our Fortran bindings and examples
works so you may want to extend the f77 deprecation period
indefinitely (or until MinGW on Wine finally can deal with Fortran 95
module files appropriately).

In sum to help figure out which of the above results are due to Wine
idiosyncracies, please let me know the -DENABLE_lua=ON results there
(assuming you actually don't have Lua installed so you will be
exercising the same CMake logic there as here); the "MSYS Makefiles"
results there for our f95 interface and examples, the version of MinGW
on your platform, and the Python results there. Finally, please
continue with the f77 deprecation including ignoring maintenance, but
put off its complete removal.

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
__________________________