Description of problem: The program runs perfectly but the output catalog has the pixel position of all detected sources fixed to 0,0 and the RA-DEC coordinates are all the same (corresponding to pixel 0,0). Version-Release number of selected component (if applicable): 2.5.0-1.fc5 How reproducible: Perfectly Steps to Reproduce: 1.Download image and sextractor input file from http://www1.cadc-ccda.hia-iha.nrc-cnrc.gc.ca/community/CFHTLS-SG/docs/cfhtlsD3.html 2.Run sextractor outputting the following parameters: X_IMAGE Y_IMAGE ALPHA_J2000 DELTA_J2000 3.Check output catalog. Actual results: All pixel positions set to 0,0 Expected results: Different pixel positions and RA-DEC's for each source. Additional info: Version directly downloaded from ftp://ftp.iap.fr/pub/from_users/bertin/sextractor/sextractor-2.5.0-1.i686.rpm works perfectly.
Thanks Pablo. The problem seems to come from the different compilation options of the home made sextractor rpm and the fedora rpm. The fedora rpm is compiled with, between other options, -O2 The rpm made by the sextractor people is compiled with -O1 As I have tested, the inclusion of -O2 of the fedora rpm produces the errors. I will tell upstream and I will produce a new rpm optimized with -O1 insted of -O2
It seems that changing the flags breaks the debugging info of the package. I have reverted the flag from -O1 to -O2 again.
Additionally, the rpm seems to work in FC4. There exists a problem with the source and gcc 4.1
Created attachment 133079 [details] Patch to fix the pointer aliasing problems