Hoi Ric,
I have your tag:
iris01>cvs diff -r rf-100902b *.*cc recoilNtp.hh
cvs diff: I know nothing about recoilDSysDict.cc
cvs diff: I know nothing about recoilNtpDict.cc
I make everything from scratch:
iris01>gmake -f Makefile clean anaRecoil
It compiles and links just fine, I run it
iris01>../bin/Linux24/anaRecoil -c chains/sx-test -y 2 -D .
Much output snipped:
./sx-test.root
Opened ./sx-test.root
*** Break *** segmentation violation
Abort
gdb shows:
(gdb) where
#0 0x40fc0091 in recoilNtp::doSplitOffStudy (this=0xbffa867c, book=true) at splitOff.icc:49
#1 0x40fcea3e in recoilNtp::bookHist (this=0xbffa867c, dump=0) at bookHist.icc:305
#2 0x0804d79e in main (argc=7, argv=0xbffff124) at anaRecoil.cc:192
#3 0x4116d647 in __libc_start_main (main=0x804bcf4 <main>, argc=7, ubp_av=0xbffff124, init=0x804b4bc <_init>,
fini=0x8063b04 <_fini>, rtld_fini=0x4000dcd4 <_dl_fini>, stack_end=0xbffff11c) at ../sysdeps/generic/libc-start.c:129
(gdb) print book
$1 = true
The problem is that *because* you have a flag to protect the booking,
the crucial return from doSplitOffStudy when called from bookHist is
not encountered. This screws Linux24. I found this out yesterday
night. I fixed it (the hard way), but you did not like my fix.
So, please provide a tag where we can run without getting debugging
trees and histograms for the splitoff study. ASAP. Else I don't think
we'll have enough time to digest the plots for Friday.
Cheers,
--U.
|