[HELP!] Athena and Artemis doesn't start
Hi to all,
This morning the "story" started with some .TRAP files during an Artemis
fit. From one of this I have understood the problem was again with conflicts
between the compiled-by-hand horae version and Carlo's .deb (automatically
installed) package.
Here the ARTEMIS.TRAP file:
# This file created at 10:32:53 on 16 April, 2004
# using linux, perl 5.008003, Tk 800.025, and Ifeffit 1.2.5
The following message was trapped by Artemis on a SIGWARN:
Artemiswarn/home/mauro/.horae/stash/ARTEMIS.TRAPCODE(0x95e0a4c) at
/usr/bin/artemis line 1491 main::__ANON__('Use of uninitialized value
in concatenation (.) or string at ...') called at /usr/bin/artemis line
5809 main::set_degeneracy(1) called at /usr/lib/perl5/Tk.pm line 225
eval {...} called at /usr/lib/perl5/Tk.pm line 225
Tk::__ANON__('Tk::Menu=HASH(0x96c8228)','active') called at
/usr/lib/perl5/Tk/Menu.pm line 531
Tk::Menu::Invoke('Tk::Menu=HASH(0x96c8228)',1) called at
/usr/lib/perl5/Tk.pm line 347 eval {...} called at /usr/lib/perl5/Tk.pm
line 347 Tk::MainLoop() called at /usr/bin/artemis line 1627
After this, I decided to reinstall everything. I have completely removed the
previous software from "/usr/local" and I have used apt-get to reinstall
horeae and ifeffit (also).
The result is: Artemis and Athena doesn't start! These the errors:
mauro@mac6npd2:~$ artemis
Can't load '/usr/lib/perl5/auto/Ifeffit/Ifeffit.so' for module Ifeffit:
/usr/lib/perl5/auto/Ifeffit/Ifeffit.so: undefined symbol: PL_sv_undef at
/usr/lib/perl/5.8/DynaLoader.pm line 225.
at /usr/bin/artemis line 49
Compilation failed in require at /usr/bin/artemis line 49.
BEGIN failed--compilation aborted at /usr/bin/artemis line 53.
mauro@mac6npd2:~$ athena
Can't load '/usr/lib/perl5/auto/Ifeffit/Ifeffit.so' for module Ifeffit:
/usr/lib/perl5/auto/Ifeffit/Ifeffit.so: undefined symbol: PL_sv_undef at
/usr/lib/perl/5.8/DynaLoader.pm line 225.
at /usr/bin/athena line 55
Compilation failed in require at /usr/bin/athena line 55.
BEGIN failed--compilation aborted at /usr/bin/athena line 59.
I really don't know what to do. I've tried different times reinstalling the
..deb package o recompiling the source.
Please, help me!
--
Mauro Rovezzi
On Friday 16 April 2004 05:51 am, mauro@rulp.org wrote:
Hi to all,
This morning the "story" started with some .TRAP files during an Artemis fit. From one of this I have understood the problem was again with conflicts between the compiled-by-hand horae version and Carlo's .deb (automatically installed) package.
This looks like a new bug that I am unaware of. I have never seen a problem in the place that the trap file is pointing to. I would be interested in understanding better what lead to this problem. If it crops up again, take notes on what you just did and let me know.
mauro@mac6npd2:~$ artemis Can't load '/usr/lib/perl5/auto/Ifeffit/Ifeffit.so' for module Ifeffit: /usr/lib/perl5/auto/Ifeffit/Ifeffit.so: undefined symbol: PL_sv_undef at /usr/lib/perl/5.8/DynaLoader.pm line 225. at /usr/bin/artemis line 49 Compilation failed in require at /usr/bin/artemis line 49. BEGIN failed--compilation aborted at /usr/bin/artemis line 53.
Whenever you see one of these error messages complaining about an undefined symbol that begins with "PL" the problem is always with PGPLOT. This could get back to your problem with having ifeffit+friends installed in different places. It would seem that either you removed pgplot entirely or pgplot is installed in the wrong place for the current setup. It may be as simple as reinstalling pgplot. Or you may need to rebuild ifeffit so that the ifeffit library is definately compiled against pgplot in its current location. Or both. Good luck! B -- Bruce Ravel ----------------------------------- ravel@phys.washington.edu Code 6134, Building 3, Room 405 Naval Research Laboratory phone: (1) 202 767 2268 Washington DC 20375, USA fax: (1) 202 767 4642 NRL Synchrotron Radiation Consortium (NRL-SRC) Beamlines X11a, X11b, X23b National Synchrotron Light Source Brookhaven National Laboratory, Upton, NY 11973 My homepage: http://feff.phys.washington.edu/~ravel EXAFS software: http://feff.phys.washington.edu/~ravel/software/exafs/
On Friday 16 April 2004 08:54 am, Bruce Ravel wrote:
This morning the "story" started with some .TRAP files during an Artemis fit. From one of this I have understood the problem was again with conflicts between the compiled-by-hand horae version and Carlo's .deb (automatically installed) package.
This looks like a new bug that I am unaware of.
Okee dokee. I am now aware of this problem. The two items in the "Set path degeneracy" submenu under the Feff menu are broken in the current version of Artemis. I don't think there is a work-around -- anyone wanting to use this feature will have to do so by hand, which is highly tedious. B -- Bruce Ravel ----------------------------------- ravel@phys.washington.edu Code 6134, Building 3, Room 405 Naval Research Laboratory phone: (1) 202 767 2268 Washington DC 20375, USA fax: (1) 202 767 4642 NRL Synchrotron Radiation Consortium (NRL-SRC) Beamlines X11a, X11b, X23b National Synchrotron Light Source Brookhaven National Laboratory, Upton, NY 11973 My homepage: http://feff.phys.washington.edu/~ravel EXAFS software: http://feff.phys.washington.edu/~ravel/software/exafs/
On Fri, Apr 16, 2004 at 09:49:55AM -0400, Bruce Ravel wrote:
On Friday 16 April 2004 08:54 am, Bruce Ravel wrote:
This morning the "story" started with some .TRAP files during an Artemis fit. From one of this I have understood the problem was again with conflicts between the compiled-by-hand horae version and Carlo's .deb (automatically installed) package.
This looks like a new bug that I am unaware of.
Okee dokee. I am now aware of this problem.
The two items in the "Set path degeneracy" submenu under the Feff menu are broken in the current version of Artemis. I don't think there is a work-around -- anyone wanting to use this feature will have to do so by hand, which is highly tedious.
B
In fact, I got this trap file while changing degeneracy from FEFF to 1.
--
Mauro Rovezzi
Mauro: Let me address the last problem. As Bruce says, I think that this is a PGPLOT inconsistency. Could you please send me the output of the following commands on your computer. You may do it directly instead of clogging the list. whereis ifeffit whereis athena locate pgplot dpkg -l | grep pgplot horae is compiled against pgplot and therefore knows where to find the libraries. If you have any mixed installation it can cause problems. In a pinch, I could walk you through compiling the packages for testing and then you would have packages custom built for your system. At the moment, I do not have a pure testing machine and until I am more familiar with the automated pbuilder scripts, I will not be able to generate them. Carlo On Fri, 16 Apr 2004, mauro@rulp.org wrote:
Hi to all,
After this, I decided to reinstall everything. I have completely removed the previous software from "/usr/local" and I have used apt-get to reinstall horeae and ifeffit (also).
The result is: Artemis and Athena doesn't start! These the errors:
mauro@mac6npd2:~$ artemis Can't load '/usr/lib/perl5/auto/Ifeffit/Ifeffit.so' for module Ifeffit: /usr/lib/perl5/auto/Ifeffit/Ifeffit.so: undefined symbol: PL_sv_undef at /usr/lib/perl/5.8/DynaLoader.pm line 225. at /usr/bin/artemis line 49 Compilation failed in require at /usr/bin/artemis line 49. BEGIN failed--compilation aborted at /usr/bin/artemis line 53.
mauro@mac6npd2:~$ athena Can't load '/usr/lib/perl5/auto/Ifeffit/Ifeffit.so' for module Ifeffit: /usr/lib/perl5/auto/Ifeffit/Ifeffit.so: undefined symbol: PL_sv_undef at /usr/lib/perl/5.8/DynaLoader.pm line 225. at /usr/bin/athena line 55 Compilation failed in require at /usr/bin/athena line 55. BEGIN failed--compilation aborted at /usr/bin/athena line 59.
I really don't know what to do. I've tried different times reinstalling the ...deb package o recompiling the source.
Please, help me!
-- Carlo U. Segre -- Professor of Physics Associate Dean for Special Projects, Graduate College Illinois Institute of Technology Voice: 312.567.3498 Fax: 312.567.3494 Carlo.Segre@iit.edu http://www.iit.edu/~segre
Bruce, Carlo, Does this:
Can't load '/usr/lib/perl5/auto/Ifeffit/Ifeffit.so' for module Ifeffit: /usr/lib/perl5/auto/Ifeffit/Ifeffit.so: undefined symbol: PL_sv_undef at /usr/lib/perl/5.8/DynaLoader.pm line 225. at /usr/bin/artemis line 49 Compilation failed in require at /usr/bin/artemis line 49. BEGIN failed--compilation aborted at /usr/bin/artemis line 53.
really imply that it's relatd to PGPLOT?? I think the PL_sv_undef error is a Perl internal code (Perl Scalar Variable), and that if Ifeffit.so could not find pgplot, the errors would be more like '_pgXXX symbol not found'. I'd guess it was a perl version conflict instead, or that Ifeffit.so was built with one perl version and used with another: this is the danger of using binary packages. If PGPLOT is built and installed with PGPLOT_install, I think this is unlikely to show up only with the Perl programs, as pgplot is statically linked to libifeffit.a. That is why that installation procedure is so strongly recommended. So, if this really is a PGPLOT issue, please make sure you're not using the PGPLOT from the debian package (for linux) or Fink package (for Mac OS X), but using the one from PGPLOT_install. Hope that helps, --Matt
Matt: Perhaps you are right. In any case, it is an issue with incompatibilities of binaries. I am working with him to fix this. In general, I am planning to set up automatic build scripts so that the build dependencies stay up to date with changes in the distribution. This is not so much a problem with the Debian stable but with the bleeding edge versions that he uses. I am trying to get these packages into the official Debian distribution where the automatic updates happen with no human intervention. Carlo On Fri, 16 Apr 2004, Matt Newville wrote:
Bruce, Carlo,
Does this:
Can't load '/usr/lib/perl5/auto/Ifeffit/Ifeffit.so' for module Ifeffit: /usr/lib/perl5/auto/Ifeffit/Ifeffit.so: undefined symbol: PL_sv_undef at /usr/lib/perl/5.8/DynaLoader.pm line 225. at /usr/bin/artemis line 49 Compilation failed in require at /usr/bin/artemis line 49. BEGIN failed--compilation aborted at /usr/bin/artemis line 53.
really imply that it's relatd to PGPLOT?? I think the PL_sv_undef error is a Perl internal code (Perl Scalar Variable), and that if Ifeffit.so could not find pgplot, the errors would be more like '_pgXXX symbol not found'. I'd guess it was a perl version conflict instead, or that Ifeffit.so was built with one perl version and used with another: this is the danger of using binary packages.
If PGPLOT is built and installed with PGPLOT_install, I think this is unlikely to show up only with the Perl programs, as pgplot is statically linked to libifeffit.a. That is why that installation procedure is so strongly recommended. So, if this really is a PGPLOT issue, please make sure you're not using the PGPLOT from the debian package (for linux) or Fink package (for Mac OS X), but using the one from PGPLOT_install.
Hope that helps,
--Matt
_______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
-- Carlo U. Segre -- Professor of Physics Associate Dean for Special Projects, Graduate College Illinois Institute of Technology Voice: 312.567.3498 Fax: 312.567.3494 Carlo.Segre@iit.edu http://www.iit.edu/~segre
On Friday 16 April 2004 10:24 am, Matt Newville wrote:
really imply that it's relatd to PGPLOT?? I think the PL_sv_undef error is a Perl internal code (Perl Scalar Variable), and that if Ifeffit.so could not find pgplot, the errors would be more like '_pgXXX symbol not found'. I'd guess it was a perl version conflict instead, or that Ifeffit.so was built with one perl version and used with another: this is the danger of using binary packages.
Matt is completely correct, both in the interpretation of the error message and of the cause. Unfortunately I answered Mauro's message before my morning cup of coffee. I saw "PL" and the "pg" neurons in my brain fired off. The risk here is relying on my brain! In the case that Mauro saw, rebuilding horae is what is called for so that the ifeffit wrapper is built against the current Ifeffit library. Sorry for the confusion. And, Mauro, I hope that are able to get back up and running without too much pain. B -- Bruce Ravel ----------------------------------- ravel@phys.washington.edu Code 6134, Building 3, Room 405 Naval Research Laboratory phone: (1) 202 767 2268 Washington DC 20375, USA fax: (1) 202 767 4642 NRL Synchrotron Radiation Consortium (NRL-SRC) Beamlines X11a, X11b, X23b National Synchrotron Light Source Brookhaven National Laboratory, Upton, NY 11973 My homepage: http://feff.phys.washington.edu/~ravel EXAFS software: http://feff.phys.washington.edu/~ravel/software/exafs/
Carlo, Bruce, At the moment I have removed the Debian package, recompiled the source and working with the "old" system. I also think that the problems are related to installing a "stable" package on "testing".
In a pinch, I could walk you through compiling the packages for testing and then you would have packages custom built for your system. At the moment, I do not have a pure testing machine and until I am more familiar with the automated pbuilder scripts, I will not be able to generate them.
Ok, no problem. I can continue working without installing the Debian
package.
We can conclude, with this experience, that the "stable" package in not 100%
compatible with a "testing" system.
Thanks, Mauro
--
Mauro Rovezzi
participants (4)
-
Bruce Ravel
-
Carlo U. Segre
-
Matt Newville
-
mauro@rulp.org