Version: 3.0.2 (using KDE KDE 3.2.1) Installed from: Compiled From Sources Most debuggers I've worked with display the true type of a C++ object pointer. It's possible that some of the code I'm looking at is compiled without RTTI but it doesn't appear the KDevelop debugger does that. Special casing instances of QObject descendant classes to use the "className()" method (as has been done for QStrings) would seem to be a reasonable work-around and a great help.
Not sure I completely understand you. Could you give a concrete example of what it shows and what you think it should show.
Ok. If in the current frame stack a pointer "foo" is declared as QObject* foo = getTheFoo(); the Watch window will show Name Type Value foo QObject* (QObject*)0x12345678 when the most derived type of the object returned by "getTheFoo()" above might actually be KDockWidget (or what have you). Not sure the best way to show this but you could put the true type in either the Type or Value column and the variable's type in the other. foo KDockWidget* (QObject*)0x12345678 or foo (QObject*) KDockWidget* (KDockWidget*)0x12345678 etc. Ideally the tree should show the members of the true type also. Maybe I'm missing something and it does this now but it doesn't appear to.
Dang. Lost my tabs. Oh well, hope you can make out what I was trying to say.
If you know how to ask gdb is a class is derived from QObject, I'm all ears. The "whatis" command will tell the exact type (say "QWidget*"), but how to do check if QWidget is derived from QObject?
apparently there's no further interest from the reporter and seemingly also very little from us.