cmake-2.4.5 was released this morning and removes a bug that I had to
workaround in a fairly ugly way for 2.4.4.  I have tested the new cmake
without the workaround, and everything seems fine.  Accordingly, I have
removed the workaround and designated 2.4.5 as the minimum version of cmake
if you are using the CVS version of PLplot.  This seems a well-tested
version of cmake which appeared rapidly after 2.4.4 (in part because of the
above bug) so I expect it will be at least several months before we see a
newer version of cmake.

Let me know if you have any problems with the PLplot build with the new
version of cmake.

The only important outstanding PLplot/CMake issues I am aware of are the
following:

* The wxwidgets device driver fails to compile on systems that use the 4.x
   version of gcc (and g++).  On that system (Ubuntu dapper) I have temporarily
   worked around the problem by specifying -DPLD_wxwidgets=OFF but ultimately
   the solution is a wxwidgets.cpp change that is more explicit about what
   overloaded form of wxString is desired.

* There is something wrong with the error reporting for ctest.  For example,
   there is still a minor issue with lena.pgm and java, but you would never
   know it from the non-verbose ctest report.  Here is the verbose output:

   Testing front-end java
   PLplot library version: 5.7.1
   File lena.pgm not found
   -- Process completed
   Passed

   Even for the --verbose case, you might not even notice that remark about
   lena.pgm and instead focus on the "Passed" result which indicates all
   is well when it isn't.

* We need a libltdl build solution for windows so that we can try that
   library on that platform for the dynamic (plug-in) form of our devices.  I
   have tracked down how KDE handled the libltdl build and use for both Unix
   and windows, but I haven't implemented our own version yet due to lack of
   time.  If one of our windows developers is interested in taking this over,
   I can give you details of the KDE lead I was following.  Otherwise, I think
   this will be after our next release.

* cmake parsing of fixed-form fortran file comments is not skipped so that
   those comment lines are searched by the cmake parser for commands such as
   "use" that affect modules.  This is cmake bug 3109.  We have worked around
   this bug (in an extremely ugly way) by replacing "use " by "use_" for the
   comment lines in all our fortran 77 interface code and examples.

* The fortran 95 module dependency process is screwed up for cmake even for
   free-form fortran files where bug 3109 is not relevant.  This is cmake bug
   3984 where I give a simple example of the problem.  We have worked around
   this cmake fortran module dependency bug by creating a file (with
   arbitrary contents) called plplot.mod.proxy in the top-level build
   directory at cmake time (see examples/f95/CMakeLists.txt.)

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); PLplot scientific plotting software
package (plplot.org); the Yorick front-end to PLplot (yplot.sf.net); the
Loads of Linux Links project (loll.sf.net); and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Plplot-devel mailing list
Plplot-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/plplot-devel

Reply via email to