After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 86339 - mini-commander doesnot have a cursor
mini-commander doesnot have a cursor
Status: VERIFIED FIXED
Product: gtk+
Classification: Platform
Component: Widget: Other
2.0.x
Other Solaris
: Normal major
: ---
Assigned To: gtk-bugs
Luis Villa
: 86980 (view as bug list)
Depends on:
Blocks:
 
 
Reported: 2002-06-24 13:46 UTC by Yogeesh
Modified: 2009-08-15 18:40 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
Patch which cuases notify handlers to be called (854 bytes, patch)
2002-07-02 11:56 UTC, padraig.obriain
none Details | Review

Description Yogeesh 2002-06-24 13:46:28 UTC
I tested with 19th June build on Sparc Solaris 5.9 64 bit.
I am not able to get the cursor on mini-commander.

steps to reproduce:
1. Add minicommander to panel by selecting 
   add to panel ->utility -> Command line.
2. mini-commander comes up on the panel.
3. type any excutable command, say 'xterm', but u never find a cursor on 
mini-commander.
Comment 1 Luis Villa 2002-06-25 16:17:03 UTC
This works fine here and has been reported repeatedly in the past and
marked fixed. [See bug 74511 and duplicates.] What are the build
details on this bug?
Comment 2 Jayaraj P R 2002-06-27 17:19:44 UTC
I'm seeing this problem and i think the problem is in gtk+. I had a 
build of the GNOME sources on Jun 5th and Jun 7th. Both the builds had 
gtk+ from CVS HEAD. On the Jun 5th build, this problem is NOT seen and 
on Jun 7th this problem is seen. So, I tested the Jun 7th build with 
only gtk+ checked out with Jun 5th tag and observed that the problem 
disappears. Hence the conclusion that the problem is actually in gtk+.

The Jun 5th build used a date tag of "Jun 05 18:00:00 GMT" for gtk+
The Jun 7th build used a date tag of "Jun 07 18:00:00 GMT" for gtk+
Comment 3 Luis Villa 2002-06-27 17:31:11 UTC
This is with gtk HEAD?
Comment 4 Jayaraj P R 2002-07-01 04:48:14 UTC
Louie: Yes, gtk+ is from HEAD.
Comment 5 Owen Taylor 2002-07-01 13:33:03 UTC
*** Bug 86980 has been marked as a duplicate of this bug. ***
Comment 6 Owen Taylor 2002-07-01 13:34:56 UTC
A standalone test case not involving the panel would be useful;
perhaps someone could try and make a patch to modify
tests/testsocket to exhibit this bug.

(I don't run GTK+ HEAD currently for my GNOME desktop, which,
among other things makes debugging something like this in 
the panel hard.)
Comment 7 padraig.obriain 2002-07-01 13:43:14 UTC
I have added the keyword keynav as bug 86980 which has been marked as
a duplicate of this bug reports that focus indication is not drawn for
out-of-process applet.
Comment 8 padraig.obriain 2002-07-02 10:14:00 UTC
I do not have a standalone test case yet but I will describe what I
have discovered while investigating the problem I described in bug
86890.

The reason that the focus indicator is not drawn around an
out-of-process PanelApplet is the GTK_WIDGET_HAS_FOCUS(Widget) is
FALSE after gtk_widget_grab_widget(widget) is called for the
PanelApplet.

The reason for this is that window->has_focus is not set for the plug
in gtk_window_real_set_focus(). It looks like window->has_focus should
be set in window_update_has_focus() but when this function is called
window->is_active is FALSE.

In testsocket window->is_active is set for the plug when
socket_update_active() is called for the socket. When
socket_update_active() is called from gtk_socket_add_window()
is_active is set for the socket's toplevel window so an
XEMBED_WINDOW_ACTIVATE message is sent to the plug.

In the panel when socket_update_active() is called for the socket the
socket's toplevel window does not have is_active set.

A quick look at the code suggests that the signal handlers defined in
gtk_socket_hierarchy_changed() should be called when the toplevel
window becomes active but it does not seem that the signal handler is
called.
Comment 9 padraig.obriain 2002-07-02 11:56:45 UTC
Created attachment 9585 [details] [review]
Patch which cuases notify handlers to be called
Comment 10 padraig.obriain 2002-07-02 12:00:35 UTC
The patch I have attached fixes the symptoms of bug 86980.
Comment 11 Jayaraj P R 2002-07-02 14:41:47 UTC
The patch above seems to fix the problem reported in this bug too.
Comment 12 Owen Taylor 2002-07-02 15:32:41 UTC
Yep, this is a bad "gotcha" in GObject. '-' and '_' can be 
used interchangeably in signals and property names, but
for the notify::PROPERTY signal + detail, the detail
needs the canonical '-'.

Please commit, thanks.
Comment 13 padraig.obriain 2002-07-02 16:26:03 UTC
Change committed
Comment 14 Yogeesh 2002-09-13 09:10:30 UTC
I verified with cvs build dated 6th sept, I am not finding this bug.
hence I am closing this.