Wireshark · Wireshark-bugs: [Wireshark-bugs] [Bug 8908] Strange behavior of tree expand/collapse in packet details

[bugzilla-daemon@xxxxxxxxxxxxx]

Jakub Zawadzki changed bug 8908

What	Removed	Added
CC		darkjames-ws@darkjames.pl

Comment # 3 on bug 8908 from Jakub Zawadzki

Fixed regression in 50479 when I find some more time I'll backport it to 1.10

Old code was using post-order traverse of tree, i.e.

expand(node)
 - call expand() for all children

 - if node need to be expanded, expand it
 - if node need to be collapsed collapse it

That's why it was working.


Still one case is left:
 - packet #1, expand TCP header, expand Flags, collapse TCP header, expand TCP
header -> Flags are not expanded

But I think it's like this in 1.8


To be honest patch for this is quite simple:

@@ -216,6 +218,11 @@ expand_tree(GtkTreeView *tree_view, GtkTreeIter *iter,
      */
     if (finfo->tree_type != -1)
        tree_expanded_set(finfo->tree_type, TRUE);
+
+    /* after expanding we might also need to expand child. */
+    g_signal_handlers_block_by_func(tree_view, expand_tree, NULL);
+    expand_finfos(tree_view, model, path, iter, FALSE);
+    g_signal_handlers_unblock_by_func(tree_view, expand_tree, NULL);
 }


But I'm not 100% sure if we want to do it.
I've tested this for XML, where almost all subtrees share the same ett number,
and it looks not so great.

---

      You are receiving this mail because:

• You are watching all bug changes.