#106 Problem with digifil

[Sophia]

Hello! When I try to use digifil to convert files from .fits to .fil, I keep getting the following error message, no matter what .fits file I use. Any help would be appreciated!

$ digifil -b 8 -d 1 -o new.fil vegas_59612_40043_J1713+0747_0002_0001.fits 

Error::stack
dsp::LoadToFil::prepare
dsp::PScrunch::prepare
Error::InvalidState
Error::message
output state has not been configured

[Willem van Straten]

Hi Sophia, I'm unable to reproduce the error with the .fits files that I have on hand. Is there any way that you can make a vegas .fits file available for download?

[Sophia]

Hi Willem. Thank you so much for your quick response! The original vegas.fits file was too large to upload, but I found another .fits file that is smaller and also gives the same error message when tying to use digifil. Please find it in the link below, and let me know if there is anything else I can do. Thanks in advance!

https://drive.google.com/file/d/1LbqUxI3A8mhub6Y61_vJIDBb4RdsT7VN/view?usp=sharing

[Willem van Straten]

Hi Sophia, Thanks for making that file available. Unfortunately, I'm still unable to reproduce the error.

Could you please double-check that you are using the latest version of dspsr? If you are, then the bug might be a memory error that depends on your operating system and compiler, which is less common and harder to debug. If this is the case, could you please try recompiling dspsr as follows?

make distclean
setenv CXXFLAGS "-g -fsanitize=address -static-libasan"
setenv CFLAGS "-g -fsanitize=address -static-libasan"

./configure
make
make install

(Adapt as necessary if your build directory is separate from your source code directory.)

... then try running again? If there is a memory fault, then dspsr might output some additional error messages that might help us to track it down.

Cheers,
Willem

[Sophia]

Hi Willem,

Thank you so much for your prompt and detailed response. It seems that the version of dspsr that was installed in the server I am using was outdated by one version, so I downloaded the latest one from the SourceForge Git repository. However, when I tried to compile it using the make command, I bumped into the following error message:

mv -f $depbase.Tpo $depbase.Plo
libtool: compile: g++ -std=c++11 -DHAVE_CONFIG_H -I. -I../.. -I../../local_include -I../../Util/epsic/src -I../../Util/epsic/src/util -I../../Util/epsic/src/util -I/opt/python/3.9/envs/pulsar/include/psrchive -I/opt/python/3.9/envs/pulsar/include/psrchive/epsic -I/opt/python/3.9/envs/pulsar/include -I/minish/svs00006/dspsr/bin/include -I/usr/local/include -fopenmp -O2 -Wall -MT Archiver.lo -MD -MP -MF .deps/Archiver.Tpo -c Archiver.C -o Archiver.o
Archiver.C: In member function 'bool dsp::Archiver::postprocess(Pulsar::Archive*)':
Archiver.C:376:57: error: 'CommandFailed' is not a member of 'Pulsar::Interpreter'
  if (interpreter->get_status() == Pulsar::Interpreter::CommandFailed)
                             ^~~~~~~~~~~~~
make[3]: *** [Makefile:714: Archiver.lo] Error 1

I edited the Archiver.C file and commented the code block that was creating this error (and crossed my fingers that it wasn't essential for the code to run), and this time dspsr complied seamlessly.

Once I was able to compile dspsr, I ran (the new version of) digifil on my PSRFits files, and this time the filterbank files are created correctly! I suspect that whoever installed dspsr in my server forgot to include 'fits' in the backends.list file when compiling it, and as a result dspsr wasn't recognizing PSRFits as a valid input format.

Once again, I tremendously appreciate your time and help, and I am eager to include the proper citation and acknowledgment to dspsr in my paper draft.

Thankfully yours, and happy New Year,

Sophia

[Willem van Straten]

Thanks for the update, Sophia (and Happy New Year!)

I'm glad to hear that dspsr is working on your end. I think that the compilation error will go away if you also update your psrchive installation, but the relevant bit of code is not vital when using only digifil.

To acknowledge dspsr in publications, please cite van Straten & Bailes (2011) https://ui.adsabs.harvard.edu/abs/2011PASA...28....1V/abstract ... thank you!