Activity for Willem van Straten
-
Willem van Straten committed [d63ca2] on Code
AT3-866 correctly restore non-linear 2-bit scale in FITSUnpacker (broken since 14 Nov 2025)
-
stop running as root in local-builder docker container
-
Merge branch 'master' of ssh://git.code.sf.net/p/dspsr/code
-
AT3-1187 unnecessary #include header files removed
-
AT3-1187 CUDA::SampleDelayEngine::set_delays verified by unit test
-
AT3-1187 CUDA::SampleDelayEngine::set_delays now correctly uses
-
AT3-1233 fix InverseFilterbank tests
-
fix CUDA compilation
-
AT3-1233 fix dspsr segfault; InverseFilterbank::make_preparations
-
After reporting the number of seconds Finished, carriage return
-
ska-pst-dspsr-builder bumped from 0.3.4 to 0.3.5
-
comply with "friendship is not declaration"
-
version changed to 2026-03-31
-
AT3-1229 updated copyright years and docstrings
-
AT3-1229 more unit tests fixed after changing the interface to the
-
AT3-1229 unit tests fixed
-
AT3-1229 InverseFilterbank instances running in different threads
-
replace assert with (more informative) throw InvalidRange Error
-
if TextInterface::Parser::process returns '-' then return unprocessed text
-
pat -b residual = scale*profile -offset - template
-
Hi Emmanuel, thank you for this additional information. I could reproduce the error on my end (tutorial failure) and have just checked in some fixes to psrdada: the psrdada library has been evolving and at some point it stopped understanding the old CPSR2 data files. This backward compatibility has been restored. Could you please git pull the latest version of psrdada and compile and install it? You might also need to recompile dspsr and install it again too.
-
dada_dbdisk backward compatible with CPSR2 data files
-
compile when MADV_POPULATE_WRITE is not defined
-
Merge branch 'master' of ssh://git.code.sf.net/p/psrdada/code
-
Merge branch 'master' of git://git.code.sf.net/p/psrdada/code
-
various small corrections
-
Merge branch 'master' of git://git.code.sf.net/p/psrdada/code
-
Merge branch 'master' of git://git.code.sf.net/p/psrdada/code
-
#include <sys/select.h>
-
Hi Emmanuel, I need to have a closer look at the first error that you reported, but the clue to the second error is ASCIIObservation: failed read HDR_VERSION It appears that your ASCII header contains only the text HDR_SIZE 4096 Does this sound correct to you? DSPSR is expecting HDR_VERSION 1.0 and the required keywords listed above: ASCIIObservation::load required keywords: TELESCOPE SOURCE FREQ BW NPOL NBIT TSAMP UTC_START OBS_OFFSET ... hope this helps!
-
verbosity reductions
-
Merge branch 'master' into weighted_mean_arrival_time
-
pat -b does not output a sensible residual
-
Fixed with commit 2aa211c8300e66514fc886698c45fc5fb82bed67 Author: Willem van Straten <vanstraten.willem@gmail.com> Date: Tue Mar 24 15:40:03 2026 +1300 bugs/510 pat -b outputs the residual of the best linear fit between the template and the observation
-
bugs/510 pat -b outputs the residual of the best linear fit between the template and the observation
-
pat -b does not output a sensible residual
-
bugs/504 FITSArchive::load_Pointing sets R.A. wrap point to 2pi
-
Telescopes::set_telescope_info converts NUPPI to Nancay (tempo2 code)
-
pat -h reports -P once
-
bug fix: pat access shift estimator only when not using MTM
-
psrsplit correctly handles files that have dispersion and/or birefringence corrected
-
fix segfault when subint has no correction history extension
-
Merge branch 'master' of ssh://git.code.sf.net/p/psrchive/code
-
Merge branch 'bugs-509-psradd-tscrunch-aux'
-
Merge branch 'issues-801-pcm-calm-overlay' into bugs-509-psradd-tscrunch-aux
-
DSPSR compile error
-
Thank you for confirming!
-
Merge branch 'at3-1184-set-obsnchan-to-pipeline-input-nchan'
-
Thanks, Ramesh! As an optimization, pkg-config is stripping /home/psr/install/include from the PSRCHIVE_CFLAGS that is passed to the dspsr build tools. However, the compiler appears to be ignoring C_INCLUDE_PATH and therefore this include path is getting missed. I've checked in a fix to the dspsr configure.ac file that should disable pkg-config from stripping system include header paths from its output, by defining PKG_CONFIG_ALLOW_SYSTEM_CFLAGS=1 before calling AX_PKG_CHECK_MODULES # ensure that...
-
Thanks, Ramesh! As an optimization, pkg-config is stripping /home/psr/install/include from the PSRCHIVE_CFLAGS that is passes to the dspsr build tools. However, the compiler appears to be ignoring C_INCLUDE_PATH and therefore this include path is getting missed. I've checked in a fix to the dspsr configure.ac file that should disable pkg-config from stripping system include header paths from its output, by defining PKG_CONFIG_ALLOW_SYSTEM_CFLAGS=1 before calling AX_PKG_CHECK_MODULES # ensure that...
-
Merge branch 'master' of ssh://git.code.sf.net/p/dspsr/code
-
bugs/133 ensure that system include paths are not stripped by pkg-config
-
bugs/133 ensure that system include paths are not stripped by pkg-config
-
Thanks, Ramesh. Could you please check if the PKG_CONFIG_SYSTEM_INCLUDE_PATH environment variable is set on your system, and if it includes /home/psr/install/include? I did the following test on my laptop, which can reproduce the "stripping" of the first -I output e$ pkg-config --cflags-only-I psrchive -I/home/willem/Pulsar/include -I/home/willem/Pulsar/include/epsic $ echo $PKG_CONFIG_SYSTEM_INCLUDE_PATH $ export PKG_CONFIG_SYSTEM_INCLUDE_PATH=/home/willem/Pulsar/include $ pkg-config --cflags-only-I...
-
AT3-1184 verbosity fix
-
AT3-1184 LoadToFITS sets OBSNCHAN to the number of channels at the time of detection
-
Merge branch 'master' into at3-1184-set-obsnchan-to-pipeline-input-nchan
-
Thanks for that, Ramesh. Stranger still $ grep PSRCHIVE_.*= Makefile PSRCHIVE_ACLOCAL = /home/psr/install/share/aclocal PSRCHIVE_CFLAGS = -I/home/psr/install/include/epsic PSRCHIVE_INCLUDE = -I/home/psr/install/include/epsic PSRCHIVE_LIBS = <blah blah blah> shows that only the epsic headers folder is included in PSRCHIVE_CFLAGS ... what is returned when you run the following command? pkg-config --cflags-only-I psrchive
-
Error running ./bootstrap
-
Excellent! Sounds like things are working. Thanks for the confirmation.
-
Hi Ramesh, that's quite strange because most of the DSPSR build depends on PSRCHIVE. It's also odd to see /home/psr/install/include/epsic on the compilation command line but not /home/psr/install/include ... Could you please also attach a copy of Kernel/Formats/fits/Makefile from your DSPSR build directory?
-
Thanks for attaching the debug output from pkg-config. It confirms that the configuration file is found in what appears to be the expected location found: /opt/prog/pulsar/lib/pkgconfig/psrchive.pc and that the version in this file is as reported added dependency [psrchive = 2025-10-31] to list With apologies for what must feel like running in circles, could you please try the following? In the PSRCHIVE source code directory (/opt/prog/psrchive), run grep AC_INIT configure.ac It should return AC_INIT([PSRCHIVE],[2026-02-26],[psrchive-developers@lists.sourceforge.net])...
-
Hi Joseph, your pkg-config question prompted me to write some long-overdue documentation of this feature: https://psrchive.sourceforge.net/pkg-config/ If you have some time after we get to the root of the problem, I'd appreciate your feedback on this documentation.
-
document use of exported pkg-config descriptor
-
remove references to Qt and CVS
-
document third-party package compilation configuration via pkg-config
-
Hi Joseph, for "Theory A" could you please run git pull in the top-level folder of the psrchive source code? (It looks like you ran this command in /opt/prog/dspsr) Also, for Theory B, you forgot to paste the output of pkg-config --debug psrchive :-)
-
Hi Joseph, Could you please test the following theories? Theory A: git fetch updates your copy of the remote repository, but it doesn't actually merge the remote copy into your local copy. You could test this by running git pull in the top-level folder of the psrchive source code; if this returns 'Already up to date.' then this theory is debunked. If it shows merging activity, then this could be the issue. Theory B: There might be two installations of psrchive on your system, and pkg-config is picking...
-
Thanks for uploading the git log output. It shows that the software was almost 8 years old, and I agree that updating dspsr is a wise thing to do, owing to changes in the libraries on which it depends (primarily PSRCHIVE). This leads to the next error. Since 2018, both DSPSR and PSRCHIVE have been updated to manage more third-party package dependencies using pkg-config. This includes the dependency between DSPSR and PSRCHIVE. You can test the version of pkg-config installed on your system with $...
-
release 2026-02-26
-
make dist bug fix: remove Signal/General from libdspsrmore SOURCES
-
adapt to breaking changes in psrchive 2026-02-26
-
release 2026-02-26
-
Merge branch 'issues-801-pcm-calm-overlay'
-
Merge branch 'bugs-509-psradd-tscrunch-aux'
-
FITSArchive::get_state_from_pol_type added
-
psradd -T does not correctly handle aux:rm or aux:dm
-
Fix merged to main branch commit 72a750f7912552294c0db13f43b64d9ea6a4ebce (HEAD -> master) Merge: 4640db25d 8d532e2c3 Author: Willem van Straten <vanstraten.willem@gmail.com> Date: Thu Feb 26 21:43:35 2026 +1300 Merge branch 'bugs-509-psradd-tscrunch-aux'
-
Hi Joseph, I checked on my end, and HAVE_MPI no longer appears in Kernel/Classes/Makefile.am. What version of DSPSR are you trying to compile? When I run git log in the top-level of the source code directory, I get commit b8de2313aefd9252ae15fec66c19ec0ea384399b (HEAD -> master, origin/master, origin/HEAD) Author: Willem van Straten <vanstraten.willem@gmail.com> Date: Fri Feb 13 09:53:47 2026 +1300 Cheers, Willem
-
issues/801 some prepare methods now operate only when new data are presented
-
Merge remote-tracking branch 'origin/master' into bugs-509-psradd-tscrunch-aux
-
issues/801 increased verbosity while chasing down bug
-
Merge branch 'master' of ssh://git.code.sf.net/p/psrchive/code
-
smint: optionally use MADM of log(error) to excise outliers
-
EstimatePlotter: optionally use MADM of log(error) to excise outliers
-
smint now also smoothes and interpolates configuration projection model parameters
-
report the average logarithm of the Hessian matrix condition for two different parameterizations
-
Merge branch 'bugs-509-psradd-tscrunch-aux' of ssh://git.code.sf.net/p/psrchive/code into bugs-509-psradd-tscrunch-aux
-
bugs/509 proj:nparam fixed (was returning nchan)
-
bugs/509 update_absolute_dispersion/rotation check absolute correction flags of each Integration
-
bugs/509 spelling fixed
-
I've implemented a fix on the bugs-509-psradd-tscrunch-aux branch. (I went with potential solution #2: have sub-integrations store their own aux:rmc flag.) I'll perform more testing before checking in to the master branch
-
bugs/509 fix Python wrapper
-
bugs/509 classes renamed for greater clarity:
-
bugs/509 ColdPlasmaMeasure state and related verbosity added for debugging purposes
-
bugs/509 Archive::init_Integration now always updates relative cold plasma corrections history
-
bugs/509 Archive::init_Integration initialises cold plasma history objects only if Integration does not already own them
-
bugs/509 responsibility for recording aux corrected state shifted to ColdPlasmaHistory children
-
bugs/509 absolute_birefringence -> absolute_rotation and verbosity added
-
psradd -T does not correctly handle aux:rm or aux:dm
1 >
