search

HiPERCAM / hipercam

Python package for handling HiPERCAM data
3 stars 4 forks source link

rtplot segmentation fault #52

Closed paulkerry1 closed 6 years ago

paulkerry1 commented 6 years ago

Using rtplot on any existing hipercam data file on an Arch Linux system causes a segmentation fault (core dumped). No plot windows appear.

hipercam and trm-pgplot are git pull updated to version as of 20Jun2018 19:00 UTC

thorin{star}14: rtplot run0019.fits first - first frame to plot [1]: ccd - CCD(s) to plot [0 for all] [0]: nx - number of panels in X [3]: bias - bias frame ['none' to ignore] [none]: defect - defect file ['none' to ignore] [none]: setup - display current hdriver window settings [False]: msub - subtract median from each window? [False]: iset - set intensity a(utomatically), d(irectly) or with p(ercentiles)? [a]: xlo - left-hand X value [-50.0]: min xhi - right-hand X value [2099.0]: max ylo - lower Y value [-8.0]: min yhi - upper Y value [1033.0]: max Segmentation fault (core dumped)

dmesg shows... [ 607.152654] rtplot[2311]: segfault at 0 ip 000014e872f20ba1 sp 00007ffd95a8ed10 error 6 in libgfortran.so.5.0.0[14e872d09000+277000]

/usr/lib/libgfortran.so.5 is from the core/gcc-libs package, which is version 8.1.1+20180531-1

trmrsh commented 6 years ago

Has it ever worked on this system? "yhi" is the last argument to be read, so beyond that point it does something. I would be fairly sure that it has to be something to do with PGPLOT given the gfortran error. Does PGPLOT itself work on the system? This is not a problem that I have on the system here so it will be a difficult one to diagnose. If you or Stu could stick in a load of print statements to narrow down where the problem occurs, that could help. At a guess, it could be an issue with my cythonised version of PGPLOT so if you could work out exactly which routine is failing and the arguments that are being sent to it (types, sizes of arrays if relevant), that would be a significant help.

paulkerry1 commented 6 years ago

I cannot replicate this issue using the same setup with the same data files on an alternative Arch Linux system, so I'm closing this.Sorry for the noise.

Cheers Paul.

Just to feedback here, pgplot was indeed the issue: an update had caused pgplot to stop working due to libraries bumping up from libgfortran.so.3 to libgfortran.so.5 - rebuilding and re-installing pgplot solved it. In case this happens again, run the pgdemo1 program bundled with pgplot to check for errors before trying rtplot from the pipeline.