[Rivet] [HZTool] Compiling an HZTool main program

[Andy Buckley]

Hannes Jung wrote:
> Hi
> 
> so, then let's remove the generator specific things from the HZTOOL lib.
> But anyway, you have to link HERWIG only if you call these subroutines,
> otherwise not....

I think that's true for the static library but not the shared one: when
using a .so/.dylib library, all symbols must be resolved at link time.

> The HBOOK error could come, if you have not specified the hbook output
> file....

I think I have done that correctly: my C++ code (attached) effectively
calls

hropen(54, "HISTO", "foo.rz", "N", 1024, istat)

and a file called foo.rz *is* created. But it doesn't contain anything.
The program locks up before hropen finishes, with the apparently
infinite table of zeros written to stderr, as in the previous logfile
that I posted.

In case HBOOK doesn't like being used from a shared lib, I've tried
building my executable with the -static flag (which anyway comes from
the cernlib script if "-safe" is not used), but then I get a missing
symbol report for hropen_... strange, since that doesn't happen with
linking against  the .so version of cernlib on my system. Has anyone got
a mini Fortran executable which uses libhztool, which I can use to check
whether my installation is ok? It would be good to have such an example
supplied with the package.

Andy
-------------- next part --------------
A non-text attachment was scrubbed...
Name: hzhepmc.cc
Type: text/x-c++src
Size: 2025 bytes
Desc: not available
Url : http://www.hepforge.org/lists-archive/rivet/attachments/20090323/dec56259/attachment.cc