Hi Jacek, On Mon, 30 Sep 2002, Jacek M. Holeczek wrote: > Hi, > > First, the "Api.h" (ACLIC) problem exists on BOTH RH6.2 and RH7.1 (in both > cases I checked with gcc295x based binaries). > On a RH6.2 with gcc-2.95.3 I use your "rh61_gcc2952" binary distribution. > On a RH7.1 with gcc-2.95.3 I use your "rh72_gcc2953" binary distribution. > Do you expect it to be o.k. ? > Yes, this must be OK. I repeat my diagnostic in my previous mail: a mismatch between libraries/LD_LIBRARY_PATH. For example, it is a frequent situation on the lxplus.cern.ch to see people running with the "rh61_gcc2952" distribution, but not realizing that the default compiler on this machine is still the old egcs1.1!! The first thing to do is to check that your compiler matches the version with which Root has been compiled. To check this, simply do g++ -v if your default g++ is 2.95, now check your LD_LIBRARY_PATH to verify that egcs1.1 libs are not referenced before the ones for gcc2.95. > Second, the "maximum line size" problem is completely unrelated to ACLIC. > It exists on RH7.1 only, but I get it with both, the old root 3.02/07 > (egcs-1.1.2 based) and new 3.03/09 (gcc2953 based). > Up to now I did not use root on RH7.1, so it may be that this problem > existed for a longer time ... . I have just realized that the simplest > thing to try is ".class TVector2". As soon as the number of lines exceeds > some "magic" number, the output is gone (it seems to me that if I do not > get the "-- Press return for more --" any more, also the whole output is > mysteriously gone - in case of TVector2 it happens between 59 and 60 > lines, in case of TVector it happens between 79 and 80 lines). > What can this be ? > I cannot reproduce this problem on any system. I also believe that this is a side-effect of the first problem above. Rene Brun
This archive was generated by hypermail 2b29 : Sat Jan 04 2003 - 23:51:12 MET