Hi Pete, On Fri, 3 Sep 2004, Peter Elmer wrote: > As you probably saw from the rest of this thread, I think it had to do > with the different compiler versions using either the static or the shared > version of the library (and the fact that the library order wasn't correct > in the link line). In any case it is fixed now. I have just tried the latest version 20040902-1035 on CERN sundev with CC 5.2 (WS6U1), 5.4 (Forte7) and 5.5 (from Sun One 8): it seems to be ok. BTW: starting with CC 5.5 the template cache (/SunWS_cache) is not used any longer, so I had to modify configure (and a few other files) to test for the CC version and make the appropriate changes. I'll send the patch in a while. Gerri > On Thu, Sep 02, 2004 at 07:08:09PM -0700, Andrew Hanushevsky wrote: > > Hi All, > > > > I recall we had this problem long time ago with Objectivity. The solution > > was to include -lCstd (or something like it). It was a problem with the > > compiler in that it didn't include the library by default (you shouldn't > > have to do that as a rule). Geri notes that his version of the compiler > > doesn't have the problem. So, I suspect that along the way we dropped > > including the library manually and it sometimes works and sometimes > > doesn't. I don't have an immediate solution to this (ther than manually > > changing the make file for the particular compiler). > > > > Andy > > > > On Thu, 2 Sep 2004, Peter Elmer wrote: > > > > > Hi Jean-Yves, > > > > > > On Thu, Sep 02, 2004 at 03:54:27PM +0200, Jean-Yves Nief wrote: > > > > I've tried it on Solaris 8 and I had a problem with some > > > > referenced symbol while I was trying to start a xrootd daemon (the > > > > problem is located in libXrdOfs.so): > > > > 040902 14:40:27 001 XrootdConfig: ld.so.1: /opt/xrd/bin/cur/xrootd: > > > > fatal: relocation error: file /opt/xrd/lib/ > > > > cur/libXrdOfs.so: symbol > > > > __1cDstd2l4n0ALchar_traits4Cc___6Frn0ANbasic_ostream4CcCTA__pkC_3_: > > > > referenced symbol > > > > not found opening shared library /opt/xrd/lib/cur/libXrdOfs.so > > > > > > > > I did not try the Solaris 9 distrib yet. > > > > > > Thanks for reporting this. I can reproduce it with the sol8 binary tarball > > > on shire01 (a sol8 machine) at SLAC. > > > > > > SLAC only uses the sol9 binary tarball, that is presumably why we didn't > > > catch it. I just verified that the sol9 binary tarball does work on a sol9 > > > machine. > > > > > > What OS version are you actually running on your machines? Can you just > > > use the sol9 build while I sort out what has happened to the sol8 one? > > > > > > Pete > > > > > > > > > > Peter Elmer wrote: > > > > > > > > > Hi All, > > > > > > > > > > Following a discussion in one of the BaBar HN fora, there is now another > > > > >new xrootd version 20040901-0956. On top of the last version > > > > >20040901-0613, this just adds a few more small fixes for building xrootd > > > > >within ROOT. If nothing else comes up, it would help BaBar test the HEAD > > > > >of ROOT CVS if this xrootd version can be added in place of the one from a > > > > >month ago. > > > > > > > > > > For the full set of changes and links to rpms/tarballs to download see > > > > > the the xrootd web page: > > > > > > > > > > http://xrootd.slac.stanford.edu > > > > > > > > > >Let us know if there are problems. > > > > > > > > > > thanks, > > > > > Pete > > > > > > > > > ------------------------------------------------------------------------- > Peter Elmer E-mail: [log in to unmask] Phone: +41 (22) 767-4644 > Address: CERN Division PPE, Bat. 32 2C-14, CH-1211 Geneva 23, Switzerland > ------------------------------------------------------------------------- >