If it was a linking problem then ls would not run at all. It is running fine and probably just prints what it gets from perror. L. 2011/9/22 Lukasz Janyst <[log in to unmask]>: > Or opendir fails and is setting errno to ELIBACC > > Lukasz > > 2011/9/22 Andrew Hanushevsky <[log in to unmask]>: >> I don't quite understand but when running this out of the script that I >> supplied in a previous mail file opendir() et. al. worked just fine. It >> didn't work if a) the library was not put in quotes, and b) the >> LD_LIBRARY_PATH wasn't set correctly. The test wworked fine on RH5. >> >> Given that one gets ELIBACC (which I was able to get as well) would indicate >> that the loader is not finding all of the dependent libraries. Well, that's >> an obvious statement. Now on to see why. Brian, could you send me your >> failing script? >> >> Andy >> >> On Thu, 22 Sep 2011, Brian Bockelman wrote: >> >>> >>> Follow-up Comment #14, bug #86984 (project xrootd): >>> >>> Hi Wei, >>> >>> XRDPOSIX_DEBUG=2 didn't give me any further debug output. >>> >>> Poking around - it seems the stuff in XrdPosixLinkage is failing, because >>> the >>> error message below is associated with the ELIBACC error code. I haven't >>> figured out any way to get at it; I'm curious what error code the call to >>> dlsym might be getting. >>> >>> Brian >>> >>> _______________________________________________________ >>> >>> Reply to this item at: >>> >>> <http://savannah.cern.ch/bugs/?86984> >>> >>> _______________________________________________ >>> Message sent via/by LCG Savannah >>> http://savannah.cern.ch/ >>> >> >