AutoSaver Issue at Startup

[greebo]

Working fine here (except for the weird issue I get whereby DarkRadiant locks up the second time I start it, something to do with the AutoSaver firing during the splash screen update, but I was getting this issue before).

The AutoSaver is firing during startup? I never had such a lockup, do you get that reliably?

[OrbWeaver]

It seems pretty reliable, every second time (and sometimes the first) I start DR it locks up permanently just after the camera window is displayed, with the splash screen still visible.

 

The stacktrace is

 

#0  0xb7f61410 in __kernel_vsyscall ()
#1  0xb738434e in writev () from /lib/tls/i686/cmov/libc.so.6
#2  0xb723bec3 in ?? () from /usr/lib/libxcb.so.1
#3  0xb723c4ab in ?? () from /usr/lib/libxcb.so.1
#4  0xb723c5df in ?? () from /usr/lib/libxcb.so.1
#5  0xb723db37 in xcb_wait_for_reply () from /usr/lib/libxcb.so.1
#6  0xb79c41eb in _XReply () from /usr/lib/libX11.so.6
#7  0xb79b2230 in XQueryPointer () from /usr/lib/libX11.so.6
#8  0xb795a686 in ?? () from /usr/lib/libgdk-x11-2.0.so.0
#9  0xb792cdfc in gdk_window_get_pointer () from /usr/lib/libgdk-x11-2.0.so.0
#10 0x0843ad7e in map::AutoMapSaver::checkSave (this=0x858d500)
at map/AutoSaver.cpp:166
#11 0x0843b70b in map::AutoMapSaver::onIntervalReached (data=0x858d500)
at map/AutoSaver.cpp:268
#12 0xb7767336 in ?? () from /usr/lib/libglib-2.0.so.0
#13 0xb7766bf8 in g_main_context_dispatch () from /usr/lib/libglib-2.0.so.0
#14 0xb7769e5e in ?? () from /usr/lib/libglib-2.0.so.0
#15 0xb776a3ac in g_main_context_iteration () from /usr/lib/libglib-2.0.so.0
#16 0xb7bbe0d4 in gtk_main_iteration () from /usr/lib/libgtk-x11-2.0.so.0
#17 0x082d2879 in process_gui () at gtkmisc.h:48
#18 0x082d21c1 in ui::Splash::queueDraw (this=0x857f420)
at ui/splash/Splash.cpp:85
#19 0x082d236b in ui::Splash::setText (this=0x857f420, text=@0xbff46c84)
at ui/splash/Splash.cpp:56
#20 0x082d238a in ui::Splash::setProgressAndText (this=0x857f420, 
text=@0xbff46c84, fraction=1) at ui/splash/Splash.cpp:69
#21 0x0820da69 in main (argc=1, argv=0xbff46d54) at main.cpp:232

 

It seems to be a problem with the gdk_window_get_pointer call, although I did try commenting out this line and the problem was not resolved (although maybe I didn't recompile properly).

[greebo]

It seems to be a problem with the gdk_window_get_pointer call, although I did try commenting out this line and the problem was not resolved (although maybe I didn't recompile properly).

Thanks, I'll have a look.

[greebo]

I changed the order a bit in AutoSaver::onIntervalReached(). The AutoSaver class is activated in MainFrame::Create() which sounds about right, but it seems there is a small timeframe between MainFrame::Create and the time the map becomes valid and/or the window manager is ready for such a call. The most recent code checks now for a valid map before querying the input states.

 

Could you please check whether my change resolves this issue?

[greebo]

Hm, according to the GTK/GDK docs, this is the function description: http://library.gnome.org/devel/gdk/stable/...dow-get-pointer

 

GdkWindow* gdk_window_get_pointer(GdkWindow *window, gint *x, gint *y, GdkModifierType *mask);

 

Obtains the current pointer position and modifier state. The position is given in coordinates relative to the upper left corner of window.

 

window : a GdkWindow

x : return location for X coordinate of pointer or NULL to not return the X coordinate

y : return location for Y coordinate of pointer or NULL to not return the Y coordinate

mask : return location for modifier mask or NULL to not return the modifier mask

Returns : the window containing the pointer (as with gdk_window_at_pointer()), or NULL if the window containing the pointer isn't known to GDK

The function call I copied from the GtkRadiant code is like this:

GdkModifierType mask;
gdk_window_get_pointer(0, 0, 0, &mask);

So what I'm wondering is, how did this ever work? It's passing a NULL pointer for the GdkWindow* argument, something which is not mentioned in the documents as valid. This code needs to be adapted a bit, I guess.

 

edit: Done, latest SVN retrieves the GdkWindow from the main window.

[OrbWeaver]

Could you please check whether my change resolves this issue?

 

Unfortunately not. Sometimes I get a stacktrace which does not include the AutoSaver stuff, and just seems to be running a gtk_main loop from Splash::queueDraw(), but I also get the same lockup on your modified version of the gdk_window_get_pointer call.

[greebo]

That's strange. Please post a stacktrace here if it seems to contain information stable enough to investigate.

[OrbWeaver]

Here's one.

 

#0  0xb7f91410 in __kernel_vsyscall ()
#1  0xb73b434e in writev () from /lib/tls/i686/cmov/libc.so.6
#2  0xb726bec3 in ?? () from /usr/lib/libxcb.so.1
#3  0xb726c4ab in ?? () from /usr/lib/libxcb.so.1
#4  0xb726c5df in ?? () from /usr/lib/libxcb.so.1
#5  0xb726db37 in xcb_wait_for_reply () from /usr/lib/libxcb.so.1
#6  0xb79f41eb in _XReply () from /usr/lib/libX11.so.6
#7  0xb79e2230 in XQueryPointer () from /usr/lib/libX11.so.6
#8  0xb798a686 in ?? () from /usr/lib/libgdk-x11-2.0.so.0
#9  0xb795cdfc in gdk_window_get_pointer () from /usr/lib/libgdk-x11-2.0.so.0
#10 0x0843add1 in map::AutoMapSaver::checkSave (this=0x858d6c0)
at map/AutoSaver.cpp:184
#11 0x0843b737 in map::AutoMapSaver::onIntervalReached (data=0x858d6c0)
at map/AutoSaver.cpp:282
#12 0xb7797336 in ?? () from /usr/lib/libglib-2.0.so.0
#13 0xb7796bf8 in g_main_context_dispatch () from /usr/lib/libglib-2.0.so.0
#14 0xb7799e5e in ?? () from /usr/lib/libglib-2.0.so.0
#15 0xb779a1e7 in g_main_loop_run () from /usr/lib/libglib-2.0.so.0
#16 0xb7bee264 in gtk_main () from /usr/lib/libgtk-x11-2.0.so.0
#17 0x0820da29 in main (argc=1, argv=0xbf9eaff4) at main.cpp:257

 

Here's another (almost useless)

 

#0  0xb7796d03 in g_main_context_dispatch () from /usr/lib/libglib-2.0.so.0
#1  0xb7799e5e in ?? () from /usr/lib/libglib-2.0.so.0
#2  0xb779a1e7 in g_main_loop_run () from /usr/lib/libglib-2.0.so.0
#3  0xb7bee264 in gtk_main () from /usr/lib/libgtk-x11-2.0.so.0
#4  0x0820da29 in main (argc=1, argv=0xbf9eaff4) at main.cpp:257

 

Third one

 

#0  0xb74f8b13 in std::string::compare () from /usr/lib/libstdc++.so.6
#1  0x08213610 in std::operator< <char, std::char_traits<char>, std::allocator<char> > (__lhs=@0x9129d40, __rhs=@0x858cf00)
at /usr/include/c++/4.2/bits/basic_string.h:2218
#2  0x0821362e in std::less<std::string>::operator() (this=0x858543c, 
__x=@0x9129d40, __y=@0x858cf00)
at /usr/include/c++/4.2/bits/stl_function.h:227
#3  0x0838e083 in std::_Rb_tree<std::string, std::pair<std::string const, boost::shared_ptr<RegisterableModule> >, std::_Select1st<std::pair<std::string const, boost::shared_ptr<RegisterableModule> > >, std::less<std::string>, std::allocator<std::pair<std::string const, boost::shared_ptr<RegisterableModule> > > >::find (this=0x858543c, __k=@0x858cf00) at /usr/include/c++/4.2/bits/stl_tree.h:1398
#4  0x0838e15c in std::map<std::string, boost::shared_ptr<RegisterableModule>, std::less<std::string>, std::allocator<std::pair<std::string const, boost::shared_ptr<RegisterableModule> > > >::find (this=0x858543c, __x=@0x858cf00)
at /usr/include/c++/4.2/bits/stl_map.h:557
#5  0x0838c7e8 in module::ModuleRegistry::getModule (this=0x8585420, 
name=@0x858cf00) at modulesystem/ModuleRegistry.cpp:175
#6  0x08437f0f in module::StaticModule<map::Map>::getModule (this=0x858cf00)
at ./modulesystem/StaticModule.h:41
#7  0x08432420 in GlobalMap () at map/Map.cpp:868
#8  0x0843ace4 in map::AutoMapSaver::checkSave (this=0x858d6c0)
at map/AutoSaver.cpp:165
#9  0x0843b737 in map::AutoMapSaver::onIntervalReached (data=0x858d6c0)
at map/AutoSaver.cpp:282
#10 0xb7797336 in ?? () from /usr/lib/libglib-2.0.so.0
#11 0xb7796bf8 in g_main_context_dispatch () from /usr/lib/libglib-2.0.so.0
#12 0xb7799e5e in ?? () from /usr/lib/libglib-2.0.so.0
#13 0xb779a1e7 in g_main_loop_run () from /usr/lib/libglib-2.0.so.0
#14 0xb7bee264 in gtk_main () from /usr/lib/libgtk-x11-2.0.so.0
#15 0x0820da29 in main (argc=1, argv=0xbf9eaff4) at main.cpp:257

 

I can't see any logic in these, it looks like it's just going round in some kind of loop. It's also really weird that I get it on the second start of DarkRadiant, not usually the first.