Wt

Hi Christian

Would the following work;

template <typename Widget>
  Widget * addWidget( std::unique_ptr< Widget > widget )
  {
    Widget * result = widget.get();
    addWidget( std::move( widget ) );
    return result;
  }

Hi Mark, thanks for looking into this.

Updating the Code in WContainerWidget.h to your suggestion throws something completely different in <tuple> in both Apps this time.

Stepping deeper into the Code, it seems there is something else:

WWebWidget *WWebWidget::parentWebWidget() const
{
  /*
   * Returns the parent webwidget, i.e. skipping composite widgets
   */
  WWidget *p = this->parent();
  while (p != nullptr && dynamic_cast<WCompositeWidget *>(p) != nullptr) // Throws here when adding to root() or setStyleClass() on root()
    p = p->parent();

  return p ? p->webWidget() : nullptr;
}

I'm still trying to locate the big differences between the working and the broken app...
the while loop goes through 2 steps before segfault

Updated to Wt 4.10.3, but still throws.

Can you post code for an actual (small) working project that can demonstrate the error?

Hello Christian,

if it would indeed not be too much trouble, a minimal example would be nice.
From a glace, seeing that the widget is already a nullptr when addWidget is called, are you certain that the widget is constructed correctly, and not destructed earlier than expected? Again, that's just from a glance.

Best,
Matthias

Thanks for looking into this.

As Expected, there is nothing wrong with Wt...

To be honest, I really don't understand how this happened.
My guess is something to do with how I handled the ConnectionPool Object for each application and the Dbo::Session Object.

I need to look further into when to use unique_ptr and when not to!

At least, now it works in both Apps, and the previous error on Session exit that let me try different things seems to also have vanished.

Thank you for your time!

Good luck in your developments!