FLTK logo

Re: [fltk.general] Re: Fl_Tabs

FLTK matrix user chat room
(using Element browser app)   FLTK gitter user chat room   GitHub FLTK Project   FLTK News RSS Feed  
  FLTK Apps      FLTK Library      Forums      Links     Login 
 All Forums  |  Back to fltk.general  ]
 
Previous Message ]New Message | Reply ]Next Message ]

Re: Re: Fl_Tabs Albrecht Schlosser Sep 02, 2021  
 
On 9/2/21 4:36 PM lifeatt...@gmail.com wrote:
>then the callback is only being called when switching to the first (n-1) tabs. I.e. if i have 4 tabs, 
>"A","B","C", and "D", the callback is only called when switching to "A", "B" or "C".

I don't see the issue.

I do. It's really strange.

I modified your example to show me which tab is active in the callback, and the callback is invoked for all tabs. (gcc 9.3.0 on Linux Mint 20).

Linux Mint 20 here as well.
$ g++ --version
g++ (Ubuntu 9.3.0-17ubuntu1~20.04) 9.3.0

I'm uncertain what your test of tabs->value() was to accomplish in the mywin handler, you will receive messages for the "previous" tab (e.g. UNFOCUS) when switching to a new tab.

Kevin, the issue is not that you will receive the value of the previous tab (which is true) but that the simple 'if ...' changes the behavior. If I'm not missing anything obvious that's a mystery (or really a C++ bug). Side note: I found another strange issue (a nonsensical compiler warning) with this g++ version. It's maybe really a bug.

I modified mywin::handle() like this:

int mywin::handle(int msg) {
int ret = Fl_Double_Window::handle(msg);
#if 1
if (tabs->value() == g1 && msg == FL_PUSH) printf("g1\n");
if (tabs->value() == g2 && msg == FL_PUSH) printf("g2\n");
if (tabs->value() == g3 && msg == FL_PUSH) printf("g3\n");
#endif
return ret;
}

If you change '#if 1' to '#if 0' the "callback behavior" changes as described by the OP in the way that "In callback" is printed only when you switch tabs to tab 1 or tab 2 but not to the last tab. That's the effect I'm seeing, and that is really unexpected. Particularly because the callback should already have been called by 'Fl_Double_Window::handle(msg);' when nthe following 3 statements are executed.

OK, I also built FLTK with clang and used debug and no-optimize (-g -O0) switches but the result is the same.

The above seems to rule out a compiler bug.

(Some time later ...)

New tests showed that calling tabs->value() appears to have a side effect that makes this issue happen. For anybody to reproduce the issue my further modified handle() looks like this:

int mywin::handle(int event) {
int ret = Fl_Double_Window::handle(event);
#if 0
void *pp = tabs->value();
// std::cerr << "tabs->value() =" << pp << std::endl;
#else
const char *pp = tabs->label() ? tabs->label() : "none";
// std::cerr << "tabs->label() = " << pp << std::endl;
#endif
return ret;
}

Note: if you comment out the output statements as shown above, make sure that the compiler doesn't optimize the calls away. It shouldn't do that with tabs->value() because it's not 'const', but anyway. Take care.

Now change '#if 0' to '#if 1' and watch the result:

(1) With '#if 0' tabs->label() is called and the code works as expected.
(2) With '#if 1' tabs->value() is called and the code fails to show "In callback" for the last tab.

BTW: I would have expected that Fl_Tabs::value() is a 'const' method but it isn't. Except returning the "current tab" value it does also show/hide tabs. This seems to have to do with this unexpected effect.

Still testing ...

--
You received this message because you are subscribed to the Google Groups "fltk.general" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fltkgeneral+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/fltkgeneral/dcd7c9e9-01a1-48de-6304-3021d2837bca%40online.de.
Direct Link to Message ]
 
     
Previous Message ]New Message | Reply ]Next Message ]
 
 

Comments are owned by the poster. All other content is copyright 1998-2024 by Bill Spitzak and others. This project is hosted by The FLTK Team. Please report site problems to 'erco@seriss.com'.