new GalSim version: v0.3

2 views
Skip to first unread message

Rachel Mandelbaum

unread,
Feb 1, 2013, 8:24:06 AM2/1/13
to galsim-...@googlegroups.com
We are pleased to announce a new tagged version of GalSim, v0.3, which can be obtained now from our GitHub page,
https://github.com/GalSim-developers/GalSim/
or from Fink.  Below is a list of changes from the previous version, reproduced from CHANGELOG.md:

* Fixed bug in config RandomCircle when using inner_radius option.

* Fixed bug in config when trying to draw objects whose postage stamp falls entirely off the main image.

* Fixed treatment of duplicate numpy.int32 types on some systems where the old check was not sufficient.

* Fix warnings in boost::random stuff on some systems (Issue #250)

* Minor changes in the python interface to the outputs of the moments and shape estimation routines (the HSMShapeData class).  (Issue #296, #316, #332)

* Support reading and writing compressed fits files. (Issue #299)

* Add FormattedStr option for string values in config files.  (Issue #315)

* On systems where a different C++ compiler was used for GalSim and for python, C++ exceptions show up with a non-informative error message.  While this is not easily fixable, there is now a test for this incompatibility when compiling, which results in a warning being generated. (Issue #317)

* Made default poisson_flux value = False when n_photons is explicitly given to drawShoot. (Issue #319)

* It is now possible to draw the Fourier images of GSObjects using the `drawK()` method, which was always available in C++ but now is visible in python as well. (Issue #319)

* Several bug fixes in the Fourier space parameters of the Sersic surface brightness profile, which improves some issues with ringing in images composed of Sersic profiles on their own or combined with other profiles. (Issues #319, #330)

* A minor fix for some issues with image types in `fits.writeCube()`. (Issue #320)

* Minor change in the keywords related to directory specification for RealGalaxyCatalog. (Issue #322)

* Fixed several sources of memory leaks, with the most significant being in the moments and shape estimation
Reply all
Reply to author
Forward
0 new messages