Alex Romosan wrote: > Package: paw > Version: 2005.05.09-3 > Severity: important > > i am trying to use ntuples but paw crashes with a segmentation > violation: > > PAW > nt/cr 1 'test' 1 ! ! x > PAW > n/pri 1 > > *** Break *** Segmentation violation > Traceq lun = 0, level = 99 > > TRACEQ. In-line trace-back still not available. > Longjump > > this also happens on another computer with paw built directly from the > cvs source from cern so i suspect the problem might be with some other > libraries (lapack is a prime suspect).
Hmm. What compiler did you use to build Paw from CERN CVS? Maybe this is a new bug in gcc 4.0. I've just tested Paw 2005.05.09-3 on Sid (where it breaks as you say) and 2004.11.04-3 on Sarge (where it works fine). Neither upstream nor I changed any of the Paw ntuple internals between those versions. Right now I am building 2005.05.09-3 on Sarge with gcc/g77 3.3 to see what happens. For what it's worth, gdb says that the breakage happens in h_load_nt() in the file pawlib/paw/ntuple/qp/qp_hbook_if.c -- looking at this file, I don't think lapack is to blame. In the meantime you ought to be able to use the version of Paw from Sarge as a workaround, although this may be annoying to install since it uses different versions of several libraries. regards, -- Kevin B. McCarty <[EMAIL PROTECTED]> Physics Department WWW: http://www.princeton.edu/~kmccarty/ Princeton University GPG: public key ID 4F83C751 Princeton, NJ 08544 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
