Hi Wim,
this is not strictly GUI issue ;-)
WLavrijsen_at_lbl.gov wrote:
> Fons,
>
>
>>What is the problem with the GUI in the interpreter?
>
>
> not wanting to cut in front of Christian, but since I seconded his
> statement, here's an example of what hit me two days ago:
>
> $ cat test.C
> #include <iostream>
>
> struct C {
> C() { std::cout << "creating a C" << std::endl; }
> ~C() { std::cout << "destroying a C" << std::endl; }
> };
>
> int test() {
> static C c;
> std::cout << "this is a little test program" << std::endl;
> return 0;
> }
> $ root -b -q -l test.C
> root [0]
> Processing test.C
>
> and then nothing ... it seems to be stuck in an infinite loop (top shows
> that root.exe takes about 45% cpu time). As an alternate, I do:
>
> $ cat test.C
> #include <iostream>
>
> struct C {
> C() { std::cout << "creating a C" << std::endl; }
> ~C() { std::cout << "destroying a C" << std::endl; }
> };
>
> static C c;
>
> int test() {
> std::cout << "this is a little test program" << std::endl;
> return 0;
> }
> $ root -b -q -l test.C
> root [0]
> Processing test.C...
> creating a C
> this is a little test program
> (int)0
>
> thus, it appears that the dtor is never called (or, if this were like VC++6,
> it could be std::cout being prematurely destroyed, but the original code had
> side effects in the dtor; so it really is the dtor not being called).
>
> Whereas ACLiC works just fine in both cases.
>
> This is on MacOSX 10.3.9, with ROOT from CVS HEAD of two days ago.
>
> What I'm trying to do in my application (running the ATLAS Athena framework
> from ROOT/CINT, then dropping into the CINT prompt and allowing people to
> interactively access their data in the Athena TDS), requires some form of
> dictionary loading and subsequent use of those classes. Hence ACLiC alone is
> not enough. However, as in the above example, CINT alone doesn't work either.
> Finding the right balance between the two is harder than I hoped it to be.
>
> Best regards,
> Wim
--
Org: CERN, European Laboratory for Particle Physics.
Mail: 1211 Geneve 23, Switzerland
E-Mail: Fons.Rademakers_at_cern.ch Phone: +41 22 7679248
WWW: http://www.rademakers.org/fons/ Fax: +41 22 7679480
Received on Tue May 17 2005 - 23:45:43 MEST