Andreas, I have made a binary distribution with the current CVS for win32gdk. Take ftp://root.cern.ch/root/root_v4.00.02a.win32gdk.tar.gz Rene Brun On Sat, 6 Mar 2004, Andreas Zoglauer wrote: > Hello, > > I got a lot of a replys - but th eproblem is still there. > > Since several persons have proposed to switch to the newest version in > cvs, but at the moment on windows I do not have the posiblity to > recompile ROOT, it is probably the best solution to wait for the new > binaries of ROOT 4.00/03 win32gdk. > > So here I give a summary of the problem: > > The idea was to try to run our ROOT based Linux software also on Windows > XP SP1 with Visual C++ 6.0 SP5 without using the cygwin environment. > I started with the binaries of ROOT v3.10/02 win32gdk (the recommended > version) and got everything to compile and link. > The one and only problem was that each call like > cout<<TString("A TString")<<endl; crashes the program. > Commenting out all those appearances, the program (consisting out of ~50 > classes) runs flawlessly and reproduces the results gained on Linux. > > I encapsulated the problem into the small program: > > #include <Riostream.h> > #include "TString.h" > int Macro() > { > cout<<"Let's crash:"<<endl; > cout<<TString("CRASH")<<endl; > return 0; > } > > Compiling this in CINT with .x Macro.C+ the program crashes in the > marked line, as it does when I compile it in Visual C++ > > Switching to the binaries of ROOT v4.00/02 win32gdk resulted in the same > behaviour. In detail, I use: "win32gdk, 4.0.2 from > WindowsXP/NT/w2000 with VC++ 7.0 (runs with VC++6), version 4.00/02 > (good old tar file) **WIN32GDK**" > > Compiling the programs in $ROOTSYS/test leaded to a crash during the > execution of the program tstring (no output was generated), but others > like event.exe run flawlessly. Here is the requested output of nmake: > > C:\Daten\Root\r040002\test>nmake -f Makefile.win32 > > Microsoft (R) Program Maintenance Utility Version 6.00.8168.0 > Copyright (C) Microsoft Corp 1988-1998. All rights reserved. > > cl -DWIN32 -D_WIN32 -D_MT -D_DLL -MD -EHsc -nologo -G5 -GR > -MD -DWIN32 > -DVISUAL_CPLUSPLUS -D_WINDOWS -IC:\Daten\Root\R040002/include -O2 -c > tstring.c > xx > tstring.cxx > link -opt:ref /NODEFAULTLIB /INCREMENTAL:NO /PDB:NONE /RELEASE > /NOLOGO - > subsystem:console,4.0 -nologo tstring.obj > C:\Daten\Root\R040002/lib/libCore.lib > C:\Daten\Root\R040002/lib/libCint.lib > C:\Daten\Root\R040002/lib/libHist.lib C: > \Daten\Root\R040002/lib/libGraf.lib > C:\Daten\Root\R040002/lib/libGraf3d.lib C:\ > Daten\Root\R040002/lib/libGpad.lib C:\Daten\Root\R040002/lib/libTree.lib > C:\Dat > en\Root\R040002/lib/libRint.lib > C:\Daten\Root\R040002/lib/libPostscript.lib C:\ > Daten\Root\R040002/lib/libMatrix.lib > C:\Daten\Root\R040002/lib/libPhysics.lib ms > vcrt.lib oldnames.lib kernel32.lib ws2_32.lib mswsock.lib advapi32.lib > user32.l > ib gdi32.lib comdlg32.lib winspool.lib MSVCPRT.LIB -out:tstring.exe > "tstring.exe done" > > Unfortunately up to now, nobody seems to be able to reproduce the > problem. ;-(( > > But there are rumours that a fix is already in cvs. So it is probably > the best idea to wait until somebody compiles a new version of win32gdk. > > Thanks and Ciao, > Andreas > > > >
This archive was generated by hypermail 2b29 : Sun Jan 02 2005 - 05:50:06 MET