From: Alan W. I. <ir...@be...> - 2003-03-31 18:48:33
|
We are now in the final stages of the release cycle for plplot-5.2.1. This version has a much more refined configuration than the 5.2.0 version (thanks mostly to Rafael Laboissiere's efforts). The infamous 5.2.0 configuration bug for the combination of static drivers and double precision is now fixed as well as many more minor configuration bugs, and we also have some new features to crow about when the time comes. Because of this configuration refinement we expect that plplot will run on a wider variety of platforms than the 5.2.0 release. For example, we have had good reports already for a limited (no octave, for example) OSF1 platform as well as a netbsd platform where more front-end options were available (no tcl/tk or java, but the C, C++, fortran, python and octave front ends were all successfully tested on netbsd). However, to make more progress in getting rid of all cross-platform portability issues we need the active cooperation of our community to test the pre-release versions of PLplot. Thus, we urge you to test the pre-release candidate that is available at http://people.debian.org/~rafael/plplot.html that will ultimately evolve into our first release candidate (scheduled to be released on April 12th with the official release date for PLplot-5.2.1 scheduled for April 19th). Warning! The tarball version at http://people.debian.org/~rafael/plplot.html is still changing on a daily basis as we discover and fix minor configuration bugs. But the two good reports we have had recently for two different unices as well the generally good reports we have for various Linux distributions indicates this is a good time to encourage the PLplot community to give the pre-release versions strong testing on as many platforms as are accessible to you. Note, when you report successes and failures please give the full-name (which includes the date) of the tarball. For example, today's version has the name plplot-5.2.0.cvs.20030331.tar.gz. Also, please describe your platform and also state the exact configuration options that you used as well as the configuration summary that is put out at the end of by ./configure. If you just have a short success report, I think it is appropriate to post it here while the more interesting discussion of bugs and their fixes should probably be posted to plplot_devel (although there are no hard and fast rules about that). To build PLplot from the tarball use the usual ./configure; make; make install. Certain aspects of the tarball have been prebuilt (as compared with the cvs version of PLplot) so you will not need autotools, swig, or docbook development tools to test and use this tarball. Note a simple test of the installed version can be made by cd $prefix/lib/plplot$version/examples; cd c; make; cd c++; make; cd f77; make; cd tk; make; cd ..; ./plplot-test.sh That last script generates a large number of example plots (postscript by default, but png, jpeg, and other file devices are available as well, see ./plplot-test.sh --help). You can also go into the c (or any other) directory of the installed examples and try some of the examples by running them interactively with the xwin, tk or other interactive devices. Note that some configuration options have changed. For example, --with-double is the default now, and to force single-precision you must specify --without-double. Also, shared libraries and dynamic drivers are the default now (in fact the python and octave front ends don't work without shared libraries), and we recommend you turn off static libraries (as an unnneeded waste of space and computer time) by using the --disable-static option. Use ./configure --help to get more extensive information about the many configuration options that are available to you. Good luck with your pre-release testing The more you do, the better this next release will be! Alan W. Irwin on behalf of the PLplot team. __________________________ Alan W. Irwin email: ir...@be... phone: 250-727-2902 Astronomical research affiliation with Department of Physics and Astronomy, University of Victoria (astrowww.phys.uvic.ca). Programming affiliations with the Canadian Centre for Climate Modelling and Analysis (www.cccma.bc.ec.gc.ca) and the PLplot scientific plotting software package (plplot.org). __________________________ Linux-powered Science __________________________ |
From: Robert S. <ro...@sc...> - 2003-04-01 07:47:48
|
On Mon, Mar 31, 2003 at 10:47:06AM -0800, Alan W. Irwin wrote: > Good luck with your pre-release testing The more you do, the better > this next release will be! ----------8<---------- robert@ganymed:~/tmp/plplot-5.2.0.cvs.20030331> ./configure [...] robert@ganymed:~/tmp/plplot-5.2.0.cvs.20030331> make [...] gcc -DHAVE_CONFIG_H -I. -I. -I../.. -I../../include -I../../include -I../../libltdl -g -O2 -mieee-fp -MT scstubs.lo -MD -MP -MF .deps/scstubs.Tpo -c scstubs.c -fPIC -DPIC -o .libs/scstubs.lo gcc -DHAVE_CONFIG_H -I. -I. -I../.. -I../../include -I../../include -I../../libltdl -g -O2 -mieee-fp -MT scstubs.lo -MD -MP -MF .deps/scstubs.Tpo -c scstubs.c -o scstubs.o >/dev/null 2>&1 mv -f .libs/scstubs.lo scstubs.lo /bin/sh ../../libtool --mode=compile -c -o strutil.lo `test -f 'strutil.f' || echo './'`strutil.f libtool: unrecognized option `-c' Try `libtool --help' for more information. make[3]: *** [strutil.lo] Error 1 make[3]: Leaving directory `/home/robert/tmp/plplot-5.2.0.cvs.20030331/bindings/f77' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory `/home/robert/tmp/plplot-5.2.0.cvs.20030331/bindings' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/home/robert/tmp/plplot-5.2.0.cvs.20030331' make: *** [all] Error 2 ----------8<---------- With --disable-f77 it compiles. Robert -- Dipl.-Ing. Robert Schwebel | http://www.pengutronix.de Pengutronix - Linux Solutions for Science and Industry Braunschweiger Str. 79, 31134 Hildesheim, Germany Handelsregister: Amtsgericht Hildesheim, HRA 2686 Phone: +49-5121-28619-0 | Fax: +49-5121-28619-4 |
From: Rafael L. <lab...@ps...> - 2003-04-02 21:42:06
|
* Robert Schwebel <ro...@sc...> [2003-04-01 09:46]: > On Mon, Mar 31, 2003 at 10:47:06AM -0800, Alan W. Irwin wrote: > > Good luck with your pre-release testing The more you do, the better > > this next release will be! > > ----------8<---------- > robert@ganymed:~/tmp/plplot-5.2.0.cvs.20030331> ./configure > [...] > robert@ganymed:~/tmp/plplot-5.2.0.cvs.20030331> make > [...] > gcc -DHAVE_CONFIG_H -I. -I. -I../.. -I../../include -I../../include -I../../libltdl -g -O2 -mieee-fp -MT scstubs.lo -MD -MP -MF .deps/scstubs.Tpo -c scstubs.c -fPIC -DPIC -o .libs/scstubs.lo > gcc -DHAVE_CONFIG_H -I. -I. -I../.. -I../../include -I../../include -I../../libltdl -g -O2 -mieee-fp -MT scstubs.lo -MD -MP -MF .deps/scstubs.Tpo -c scstubs.c -o scstubs.o >/dev/null 2>&1 > mv -f .libs/scstubs.lo scstubs.lo > /bin/sh ../../libtool --mode=compile -c -o strutil.lo `test -f 'strutil.f' || echo './'`strutil.f > libtool: unrecognized option `-c' > Try `libtool --help' for more information. > make[3]: *** [strutil.lo] Error 1 > make[3]: Leaving directory `/home/robert/tmp/plplot-5.2.0.cvs.20030331/bindings/f77' > make[2]: *** [all-recursive] Error 1 > make[2]: Leaving directory `/home/robert/tmp/plplot-5.2.0.cvs.20030331/bindings' > make[1]: *** [all-recursive] Error 1 > make[1]: Leaving directory `/home/robert/tmp/plplot-5.2.0.cvs.20030331' > make: *** [all] Error 2 > ----------8<---------- > > With --disable-f77 it compiles. Thank you for this valuable bug report. Indeed, our configure logic regarding the Fortran 77 compiler is broken, since it proceeds happily even if no compiler is found. I am going to fix this in CVS. In the meantime, since you are using Debian, configuring with --enable-f77 should work for you after "apt-get install g77". -- Rafael |
From: Robert S. <ro...@sc...> - 2003-04-01 07:54:33
|
On Mon, Mar 31, 2003 at 10:47:06AM -0800, Alan W. Irwin wrote: > Good luck with your pre-release testing The more you do, the better > this next release will be! ----------8<---------- robert@ganymed:~/tmp/plplot-5.2.0.cvs.20030331> ./configure --disable-f77 [...] robert@ganymed:~/tmp/plplot-5.2.0.cvs.20030331> make [...] robert@ganymed:~/tmp/plplot-5.2.0.cvs.20030331> cd examples/c robert@ganymed:~/tmp/plplot-5.2.0.cvs.20030331/examples/c> ./x01c *** PLPLOT ERROR *** plInitDispatchTable: Could not open drivers directory, aborting operation Segmentation fault ----------8<---------- System is debian unstable. Robert -- Dipl.-Ing. Robert Schwebel | http://www.pengutronix.de Pengutronix - Linux Solutions for Science and Industry Braunschweiger Str. 79, 31134 Hildesheim, Germany Handelsregister: Amtsgericht Hildesheim, HRA 2686 Phone: +49-5121-28619-0 | Fax: +49-5121-28619-4 |
From: Maurice L. <mj...@ga...> - 2003-04-02 20:58:44
|
Robert Schwebel writes: > On Mon, Mar 31, 2003 at 10:47:06AM -0800, Alan W. Irwin wrote: > > Good luck with your pre-release testing The more you do, the better > > this next release will be! > > ----------8<---------- > robert@ganymed:~/tmp/plplot-5.2.0.cvs.20030331> ./configure --disable-f77 > [...] > robert@ganymed:~/tmp/plplot-5.2.0.cvs.20030331> make > [...] > robert@ganymed:~/tmp/plplot-5.2.0.cvs.20030331> cd examples/c > robert@ganymed:~/tmp/plplot-5.2.0.cvs.20030331/examples/c> ./x01c > > *** PLPLOT ERROR *** > plInitDispatchTable: Could not open drivers directory, aborting operation > Segmentation fault > ----------8<---------- I think the problem is that you haven't done a 'make install' yet. Unfortunately a limitation of the current system is that it won't work until installed. Perhaps this limitation will be removed eventually to some extent. But as it stands after building you need to: make install cd <prefix>/lib/plplot<version>/examples/c make x01c ./x01c etc -- Maurice LeBrun mj...@ga... Research Organization for Information Science and Technology of Japan (RIST) |
From: Robert S. <ro...@sc...> - 2003-04-03 06:37:15
|
On Wed, Apr 02, 2003 at 02:56:25PM -0600, Maurice LeBrun wrote: > Unfortunately a limitation of the current system is that it won't work until > installed. Perhaps this limitation will be removed eventually to some extent. > But as it stands after building you need to: > > make install > cd <prefix>/lib/plplot<version>/examples/c > make x01c > ./x01c Hmm, wasn't that why there is a PLPLOT_HOME variable? Normally, from a user's point of view, I would like to be able to test a packet locally before I spread it's files over my harddisk ;) Robert -- Dipl.-Ing. Robert Schwebel | http://www.pengutronix.de Pengutronix - Linux Solutions for Science and Industry Braunschweiger Str. 79, 31134 Hildesheim, Germany Handelsregister: Amtsgericht Hildesheim, HRA 2686 Phone: +49-5121-28619-0 | Fax: +49-5121-28619-4 |
From: Rafael L. <lab...@ps...> - 2003-04-03 13:08:12
|
* Robert Schwebel <ro...@sc...> [2003-04-03 08:36]: > On Wed, Apr 02, 2003 at 02:56:25PM -0600, Maurice LeBrun wrote: > > Unfortunately a limitation of the current system is that it won't work until > > installed. Perhaps this limitation will be removed eventually to some extent. > > But as it stands after building you need to: > > > > make install > > cd <prefix>/lib/plplot<version>/examples/c > > make x01c > > ./x01c > > Hmm, wasn't that why there is a PLPLOT_HOME variable? Normally, from a > user's point of view, I would like to be able to test a packet locally > before I spread it's files over my harddisk ;) You are fully right. The present situation is unsatisfactory, and there is a consensus among the developers that it must be improved. In the past, we had a different paradigma for building the software, with all the files being put under a tmp/ directory. We switched to a autoconf/ automake/libtool based approach, which is expected to be superior in many regards, but we have lost that test-without-install feature. Well, not quite. You can try this: cd examples/c make x01c PLPLOT_DRV_DIR=../../drivers PLPLOT_LIB=../../data ./x01c This should work with the tarball that you downloaded (this is not documented, sorry). Otherwise, you can safely install PLplot after "./configure --prefix=/some/dir ; make ; make install", and than later "rm -rf /some/dir" to get rid of all the "files over the harddisk". At any rate, a clean implementation of this test-without-install feature is in our agenda, but we prefered not to implement this for the forthcoming release, because it is expected to be just a bug-fixing release after 5.2.0. -- Rafael |