H.-Gerd Rosarius, 0. look at test/guiviewer.h test/guivieer.cxx code and Makefile 1. use TApplication (not TGApplication). 2. do not include RQ_OBJECT( "MFTest" ) That is required anly for interpreted classes. 3. create LinkDef.h and dictionary for MFTest class Must work. HTH. Regards. Valeriy > Hey Rene, > > thanks for your explanation about the latest development in windows GUI > matters. Now I have a clear picture! > > However, I would like to use the chance and show you my code that should > work with the ROOT win32gdk-release for windows. (See below) Maybe I can > get it to work with yours and the lists help. :o) > > As I wrote it simply doesn't work. It crashes at runtime when reaching > MapWindow() in the constructor of my class MFTest. > It works fine if I compile it without the ClassDef macro and ignore the > dictionary. Because of the GUI event handling which I need later > I cannot ignore the dictionary. By the way: rootcint does not complain > while I create the dictionary. > > Any ideas? > > Tech: Using ROOT 3.05/07 (Win32gdk) on Win2000 with MS VC++6 installed. > Problem happend with binary and self-compiled release. > > Thanks in advance > > H.-Gerd Rosarius > > > > file://- Main -------------------------------------------------------------- > // Root-Libraries > #include <TGApplication.h> > #include <TGClient.h> > > // Own class > #include "MFTest.h" > > int main ( int argc, char** argv ) > { > TGApplication application( "ROOT GUI in Windows 2000", &argc, argv ); > MFTest t( gClient->GetRoot() ); > application.Run(); > return 0; > } > > > file://- MainFrame-Test Header ------------------------------------------- > #ifndef MFTEST_H > #define MFTEST_H > > #include <TGFrame.h> > #include <RQ_OBJECT.h> > using namespace std; > > class MFTest : public TGMainFrame > { > RQ_OBJECT( "MFTest" ) > > public: > MFTest( const TGWindow*, UInt_t = 800, UInt_t = 600 ); > ~MFTest(); > > // Macro > ClassDef( MFTest, 1 ); > }; > #endif > > file://- MainFrame-Test Implementation ----------------------------------- > #include "MFTest.h" > > MFTest::MFTest( const TGWindow* p, UInt_t height, UInt_t width ) : > TGMainFrame( p, height, width ) > { > // Causes crash > MapWindow(); > } > > MFTest::~MFTest() {} > > > > > On Thu, 2003-07-31 at 09:36, Rene Brun wrote: > > Hi, > > > > Unfortunately, you must have been very confused by Valeri Fine's > > posting. His version of ROOT with Qt is not yet ready for > > distribution. It is not yet fully compliant with the TVirtualX > > interface. Valeri's version will be incorporated into the ROOT > > official distribution only in a few weeks from now. > > > > The only interface fully TVirtualX compliant is the win32gdk version. > > Please download this version from the Web site and nothing else. > > The win32gdk version in release 3.05/07 is still a slow version. > > We had no time to include a much faster version before the release, > > but it will be available before end of August. > > > > Rene Brun > > > > On 31 > > Jul > > 2003, > > H.-Gerd Rosarius wrote: > > > > > Hello Rooters, > > > > > > after a lot of roottalk reading and unsuccessful coding I am > > > quite confused as well as frustrated about ROOT's ability to create > > > GUI applications for windows. > > > > > > My mission is simple: Creating a standalone GUI application that > > > has no other functionality than to display ROOT histograms and is > > > portable between Windows and Linux. (while Linux isn't the problem) > > > The histograms are serialized in a root-file. > > > > > > First I tried to build such an application for MS Windows 2000 with > > > the ROOT TG*-Classes. To cut a long story short: I didn't manage > > > to do it, because the application always crashed when I used > > > a simple self-written class that was derived from TGMainFrame, > > > included the RQ_OBJECT- and ClassDef-Makro. It goes without saying > > > that I made a dictionary, too. > > > > > > However, another attempt was to build a QT-Application that used > > > the TQtWidget-class out of the "ROOT Qt-Edition" from BNL. > > > (See http://root.cern.ch/root/roottalk/roottalk03/3194.html and > > > http://root.cern.ch/root/roottalk/roottalk03/3197.html) > > > > > > Valeri wrote back that this won't work. > > > (See http://root.cern.ch/root/roottalk/roottalk03/3203.html) > > > > > > Just a few minutes ago I read this posting from 1999. > > > (http://root.cern.ch/root/roottalk/roottalk99/0491.html) > > > Fons wrote on a similar task that "this should be not much of a > > > problem". But on a first glance the posting seems heavily > > > Linux related, but I'll try it in this way next. > > > > > > But please, could you clear my picture about the status of Windows > > > GUI and the different GUI related ROOT "distributions" (Qt-Edition, > > > win32gdk, etc.)?? > > > > > > Greetings > > > > > > H.-Gerd Rosarius > > >
This archive was generated by hypermail 2b29 : Thu Jan 01 2004 - 17:50:14 MET