Re: running without a mainlopp for unittests

[Stefan Kost <ensonic hora-obscura de>]

I am just not getting it to work.

Funny thing is it works in NOFORK mode, but segfaults in FORK mode.
what I do different is:
  - setting tp the GUI object
  - call gtk_show_all(window)
  - back in unittest do
    gtk_widget_destroy(window)
    while(gtk_event_pending()) gtk_main_iteration();
  - main_level is 0 all the time

I'll now try the timeout thing again...

Stefan

Tristan Van Berkom wrote:
> On Thu, 03 Feb 2005 20:41:39 +0100, Stefan Kost
> <kost imn htwk-leipzig de> wrote:
>
> > Hi,
> >
> > I use check too ;).
> > Are your tests available somehwere ?
>
>
> Hmmm, like I said; I dont really use them for GUI elements so my
> mainloop tests dont include gtk_init()... the test bodies that use mainloops
> have this kind of flow:
>
>  - create mainloop
>  - provoke something that will happen in the event queue;
>    the example I was just looking at is a test for a system wide
>    observer model that uses an IPC; one test is that I modify
>    a subject that I'm watching (through a callback that is obscurly
>    registered through a GIOChannel)
> -  register an idle func to terminate the main loop
>  - If my message was recieved properly in the related callback, I set
>    a global variable "success" to TRUE
> - call g_main_loop_run()
> - fail_unless (success == TRUE, "message");
>
> Cheers,
>                                        -Tristan