[Daqpp-forum] vmedaq analysis on 64bit ...

[Matej <mat...@ij...>]

I have tried to compile vmedaq on 64bit machine and it fails...

The problem lies in (output from make command)

make[3]: Entering directory `/opt/brachy/daqpp-cvs/vmedaq/gui'
RunMenu.cc: In function =E2=80=98gboolean do_send_command(Runmenu*)=E2=80=
=99:
RunMenu.cc:32: error: cast from =E2=80=98void*=E2=80=99 to =E2=80=98int=E2=
=80=99 loses precision

make[3]: Entering directory `/opt/brachy/daqpp-cvs/vmedaq/modules'
MyModuleGuiConfig.cc: In static member function 'static void=20
MyModuleGuiConfig::onBtnToggled(GtkWidget*, MyModuleGuiConfig*)':
MyModuleGuiConfig.cc:269: error: cast from 'void*' to 'int' loses precision
MyModuleGuiConfig.cc:270: error: cast from 'void*' to 'int' loses precision
MyModuleGuiConfig.cc: In static member function 'static void=20
MyModuleGuiConfig::onTxtChanged(GtkWidget*, MyModuleGuiConfig*)':
MyModuleGuiConfig.cc:345: error: cast from 'void*' to 'int' loses precision


These errors are solved with casting to long instead of int:
//mat...@ij...
//int thing =3D (int)ptr_to_sth;
long thing =3D (long)ptr_to_sth;

So, compiled code seems to work somehow. But simple observer class with onl=
y=20
DataRecord implemented, where some variable is increased on each new data=20
record (so data counting observer) tells me that there are exactly 5000 dat=
a=20
records in the file with 10000 data records. The is to be expected, since o=
n=20
32bit machine:
size of int  =3D 4
size of long =3D 4

but on 64bit:
size of int  =3D 4
size of long =3D 8

I expect that everything would work if the data were taken on 64bit machine=
=20
and then analyzed on 64bit as well, but I have 32bit machine in my setup, b=
ut=20
I would like to use (brand new and very fast :-) ) 4-processor 64bit comput=
er=20
for analysis. I also would not like to use 32bit precompiled binaries on=20
64bit, since I use shared libraries (ROOT mostly).=20

So, any suggestion will be highly appreciated and forever praised :-)

Best, etc,
Matej