ANNOUNCEMENT: Live Wireshark University & Allegro Packets online APAC Wireshark Training Session
April 17th, 2024 | 14:30-16:00 SGT (UTC+8) | Online

Wireshark-dev: Re: [Wireshark-dev] [Wireshark-commits] rev 38546: /trunk/gtk/ /trunk/gtk/: foll

From: Bill Meier <wmeier@xxxxxxxxxxx>
Date: Mon, 15 Aug 2011 13:47:15 -0400
On 8/15/2011 12:21 PM, Stephen Fisher wrote:
This call:

    gtk_quit_add_destroy(gtk_main_level(), GTK_OBJECT(streamwindow));

Appears to tell GTK to destroy the object "streamwindow" if Wireshark is
being closed entirely.  This then calls the signal that was connected a
few lines earlier:

   g_signal_connect(streamwindow, "destroy", G_CALLBACK(follow_destroy_cb), NULL);

Which calls the function "follow_destroy_cb" when the object
"streamwindow" is destroyed.  That "follow_destroy_cb" deletes the
temporary file used by the Follow TCP Stream feature (the Follow TCP
feeds the data into a temp file whereas Follow UDP and Follow SSL both
feed it via a tap).

So I think something needs to be substitued if they've gotten rid rid of
gtk_quit_add_destroy() in GTK 3..



I agree. I was looking at this a couple of days ago and ran a simple test.

Exiting Wireshark (with no gtk_quit_destroy()) in the follow_stream code left a file in the temp directory.

I was thinking about using "destroy_with_parent" but but then I starting going down the rathole about modal windows and etc and having to re-remember details about how GTK works.


There's also some discussion on the net about gtk_quit.... (which I didn't really understand) about the difficulties of substituting other code for this feature.

In any case, I ran out of time (and patience) to dig in further.

(I did remove two instances of the use of gtk_quit_destroy() in SVN #38415 which AFAIKT weren't needed since there was no special action taken when the window was destroyed).

Bill