Losing the keyboard on F12

[Getafix]

Replicated using Fedora 22 x86_64 live usb on a Lenovo T430 Intel HD4000 laptop!

Code: Select all

[liveuser@localhost ~]$ uname -rvp
4.0.4-301.fc22.x86_64 #1 SMP Thu May 21 13:10:33 UTC 2015 x86_64
[liveuser@localhost ~]$ wmctrl -m
Name: GNOME Shell
Class: N/A
PID: N/A
Window manager's "showing the desktop" mode: N/A

EDIT #1: Added Latest.log with display.initGL = yes.
At the end of the log the fast switching is obvious.
When the sequence ends it's because I lost the keyboard.
Nothing in the log indicates that I lost the keyboard.

Code: Select all

Opening log for Oolite version 1.82 (x86-64 test release) under Linux 4.0.4-301.fc22.x86_64 at 2015-05-31 14:17:52 -0400.
4 processors detected.
Build options: OpenAL, new planets, JavaScript console support, OXP verifier, localization tools, debug GraphViz support, JavaScript profiling.

Note that the contents of the log file can be adjusted by editing logcontrol.plist.

14:17:52.885 [dataCache.notFound]: No data cache found, starting from scratch.
14:17:52.908 [display.initGL]: V-Sync requested.
14:17:52.908 [display.initGL]: ----- WARNING: Could not enable V-Sync. Please check that your graphics driver supports the [GLX_SGI/GLX_MESA]_swap_control extension.
14:17:52.908 [display.mode.list.native]: X11 native resolution detected: 1600 x 900
14:17:52.908 [display.initGL]: Trying 32-bit depth buffer
14:17:52.927 [display.initGL]: Trying 24-bit depth buffer
14:17:52.968 [joystick.init]: Number of joysticks detected: 0
14:17:52.968 [rendering.opengl.version]: OpenGL renderer version: 3.0.0 ("3.0 Mesa 10.5.4"). Vendor: "Intel Open Source Technology Center". Renderer: "Mesa DRI Intel(R) Ivybridge Mobile ".
14:17:52.968 [rendering.opengl.extensions]: OpenGL extensions (204):
GL_ARB_draw_instanced, GL_ARB_occlusion_query, GL_ARB_texture_storage, GL_ARB_sync, GL_ARB_texture_env_dot3, GL_EXT_texture_env_add, GL_ARB_buffer_storage, GL_ARB_texture_multisample, GL_ARB_explicit_uniform_location, GL_KHR_context_flush_control, GL_NV_texture_env_combine4, GL_ARB_point_parameters, GL_ARB_vertex_array_object, GL_ARB_blend_func_extended, GL_ARB_seamless_cubemap_per_texture, GL_EXT_provoking_vertex, GL_ARB_compressed_texture_pixel_storage, GL_EXT_blend_equation_separate, GL_EXT_texture_snorm, GL_ARB_texture_non_power_of_two, GL_ARB_texture_mirrored_repeat, GL_ARB_debug_output, GL_EXT_multi_draw_arrays, GL_ARB_timer_query, GL_SGIS_texture_edge_clamp, GL_ARB_vertex_program, GL_EXT_abgr, GL_ARB_draw_buffers_blend, GL_ATI_texture_float, GL_ARB_shadow, GL_ARB_shader_atomic_counters, GL_EXT_texture_compression_dxt1, GL_EXT_texture_cube_map, GL_NV_light_max_exponent, GL_NV_primitive_restart, GL_EXT_framebuffer_object, GL_EXT_texture_sRGB, GL_MESA_pack_invert, GL_3DFX_texture_compression_FXT1, GL_ARB_conservative_depth, GL_ARB_get_program_binary, GL_ARB_clear_texture, GL_ARB_multitexture, GL_AMD_seamless_cubemap_per_texture, GL_SUN_multi_draw_arrays, GL_EXT_texture_compression_rgtc, GL_ARB_shader_objects, GL_ARB_pixel_buffer_object, GL_EXT_blend_func_separate, GL_EXT_bgra, GL_ARB_draw_buffers, GL_APPLE_packed_pixels, GL_ARB_texture_env_crossbar, GL_NV_texgen_reflection, GL_ARB_fragment_shader, GL_EXT_texture_swizzle, GL_EXT_fog_coord, GL_EXT_vertex_array_bgra, GL_EXT_framebuffer_blit, GL_ARB_texture_compression_rgtc, GL_EXT_transform_feedback, GL_OES_read_format, GL_ARB_sample_shading, GL_ARB_seamless_cube_map, GL_ARB_shading_language_420pack, GL_EXT_texture_shared_exponent, GL_ATI_blend_equation_separate, GL_APPLE_object_purgeable, GL_EXT_polygon_offset, GL_INGR_blend_func_separate, GL_EXT_compiled_vertex_array, GL_KHR_debug, GL_ARB_copy_image, GL_AMD_shader_trinary_minmax, GL_APPLE_vertex_array_object, GL_ARB_vertex_type_2_10_10_10_rev, GL_EXT_separate_specular_color, GL_ARB_shading_language_100, GL_EXT_vertex_array, GL_EXT_framebuffer_multisample_blit_scaled, GL_ARB_texture_float, GL_EXT_rescale_normal, GL_ARB_texture_swizzle, GL_EXT_packed_depth_stencil, GL_ARB_half_float_pixel, GL_EXT_draw_instanced, GL_AMD_draw_buffers_blend, GL_ARB_occlusion_query2, GL_ARB_texture_rgb10_a2ui, GL_SGIS_texture_lod, GL_ARB_derivative_control, GL_ARB_framebuffer_object, GL_AMD_conservative_depth, GL_NV_conditional_render, GL_ARB_texture_rg, GL_EXT_stencil_two_side, GL_ARB_color_buffer_float, GL_EXT_blend_color, GL_EXT_texture_rectangle, GL_NV_packed_depth_stencil, GL_ARB_vertex_array_bgra, GL_NV_texture_rectangle, GL_ANGLE_texture_compression_dxt3, GL_EXT_texture_integer, GL_EXT_texture_array, GL_ANGLE_texture_compression_dxt5, GL_ARB_shading_language_packing, GL_ARB_fragment_program_shadow, GL_ARB_map_buffer_range, GL_ARB_copy_buffer, GL_ARB_fragment_program, GL_EXT_texture, GL_S3_s3tc, GL_NV_depth_clamp, GL_EXT_draw_buffers2, GL_EXT_texture_lod_bias, GL_EXT_texture_filter_anisotropic, GL_ARB_transform_feedback2, GL_EXT_secondary_color, GL_ARB_ES3_compatibility, GL_ARB_transform_feedback3, GL_EXT_framebuffer_multisample, GL_ARB_window_pos, GL_EXT_texture_env_combine, GL_IBM_rasterpos_clip, GL_ARB_texture_compression_bptc, GL_ARB_vertex_shader, GL_SGIS_generate_mipmap, GL_ARB_explicit_attrib_location, GL_EXT_packed_float, GL_ARB_texture_border_clamp, GL_ARB_internalformat_query, GL_NV_blend_square, GL_EXT_texture3D, GL_ARB_ES2_compatibility, GL_ARB_texture_query_lod, GL_ARB_invalidate_subdata, GL_EXT_texture_sRGB_decode, GL_ARB_texture_compression, GL_ARB_texture_env_add, GL_ARB_provoking_vertex, GL_ARB_texture_rectangle, GL_EXT_texture_edge_clamp, GL_MESA_texture_signed_rgba, GL_ARB_draw_elements_base_vertex, GL_ARB_fragment_coord_conventions, GL_ARB_separate_shader_objects, GL_ARB_texture_cube_map, GL_IBM_multimode_draw_arrays, GL_ARB_vertex_attrib_binding, GL_ARB_clear_buffer_object, GL_EXT_framebuffer_sRGB, GL_ARB_vertex_type_10f_11f_11f_rev, GL_ARB_robustness, GL_ARB_point_sprite, GL_INTEL_performance_query, GL_ARB_texture_gather, GL_SGIS_texture_border_clamp, GL_ARB_depth_texture, GL_EXT_texture_object, GL_ARB_instanced_arrays, GL_EXT_blend_minmax, GL_ATI_draw_buffers, GL_ARB_conditional_render_inverted, GL_ARB_multi_bind, GL_OES_EGL_image, GL_ARB_framebuffer_sRGB, GL_EXT_shader_integer_mix, GL_EXT_blend_subtract, GL_EXT_packed_pixels, GL_IBM_texture_mirrored_repeat, GL_ARB_texture_mirror_clamp_to_edge, GL_ARB_texture_cube_map_array, GL_EXT_draw_range_elements, GL_ARB_sampler_objects, GL_ARB_half_float_vertex, GL_ARB_depth_clamp, GL_ARB_texture_storage_multisample, GL_ARB_transpose_matrix, GL_ARB_texture_query_levels, GL_ARB_vertex_buffer_object, GL_EXT_pixel_buffer_object, GL_ARB_transform_feedback_instanced, GL_ARB_texture_view, GL_ATI_texture_env_combine3, GL_EXT_polygon_offset_clamp, GL_ARB_texture_env_combine, GL_ARB_shader_bit_encoding, GL_ARB_shader_texture_lod, GL_EXT_timer_query, GL_EXT_gpu_program_parameters, GL_ARB_uniform_buffer_object, GL_ARB_multisample, GL_EXT_texture_env_dot3, GL_EXT_shadow_funcs, GL_ARB_depth_buffer_float, GL_ARB_map_buffer_alignment, GL_EXT_copy_texture, GL_EXT_stencil_wrap, GL_EXT_point_parameters, GL_ATI_separate_stencil, GL_AMD_performance_monitor, GL_EXT_subtexture, GL_MESA_window_pos
14:17:52.973 [rendering.opengl.shader.support]: Shaders are supported.
14:17:52.973 [dataCache.notFound]: No data cache found, starting from scratch.
14:17:52.974 [searchPaths.dumpAll]: Resource paths: 
    ~/GNUstep/Applications/Oolite/oolite.app/Resources
    AddOns
    AddOns/Basic-debug.oxp
14:17:53.112 [shipData.load.begin]: Loading ship data.
14:17:53.574 [script.javascript.init]: JavaScript reset successful.
14:17:53.990 [debugTCP.disconnect]: No connection to debug console: "Connection to debug console failed: 'NSPOSIXErrorDomain 111' (outStream status: 7, inStream status: 7)."
14:17:53.990 [debugTCP.send.warning]: Error sending packet header, retrying.
14:17:54.014 [debugTCP.send.error]: The following packet could not be sent: {"Oolite version" = 1.82; "packet type" = "Request Connection"; "protocol version" = 65792; }
14:17:54.018 [debugTCP.disconnect]: No connection to debug console: "Connection to debug console failed: 'bad stream.' (outStream status: 0, inStream status: 0)."
14:17:54.018 [debugTCP.connect.failed]: Failed to connect to debug console at address 127.0.0.1:8563.
14:17:54.030 [display.initGL]: Requested a new surface of 800 x 600, windowed.
14:17:54.030 [display.initGL]: Created a new surface of 800 x 600, windowed.
14:17:54.064 [startup.complete]: ========== Loading complete in 1.18 seconds. ==========
14:17:57.691 [gnustep]: 2015-05-31 14:17:57.691 oolite[2676] Defaults path '/home/liveuser/GNUstep/Defaults' did not exist - created it

14:17:57.691 [gnustep]: 2015-05-31 14:17:57.691 oolite[2676] Creating empty user defaults database

14:17:57.692 [display.initGL]: Requested a new surface of 1600 x 900, fullscreen.
14:17:57.712 [display.initGL]: Created a new surface of 1600 x 900, fullscreen.
14:17:57.842 [display.initGL]: Requested a new surface of 800 x 600, windowed.
14:17:57.874 [display.initGL]: Created a new surface of 800 x 600, windowed.
14:17:58.008 [display.initGL]: Requested a new surface of 1600 x 900, fullscreen.
14:17:58.010 [display.initGL]: Created a new surface of 1600 x 900, fullscreen.
14:17:58.141 [display.initGL]: Requested a new surface of 800 x 600, windowed.
14:17:58.173 [display.initGL]: Created a new surface of 800 x 600, windowed.
14:17:58.291 [display.initGL]: Requested a new surface of 1600 x 900, fullscreen.
14:17:58.296 [display.initGL]: Created a new surface of 1600 x 900, fullscreen.
14:17:58.424 [display.initGL]: Requested a new surface of 800 x 600, windowed.
14:17:58.457 [display.initGL]: Created a new surface of 800 x 600, windowed.
14:17:58.557 [display.initGL]: Requested a new surface of 1600 x 900, fullscreen.
14:17:58.563 [display.initGL]: Created a new surface of 1600 x 900, fullscreen.
14:17:58.674 [display.initGL]: Requested a new surface of 800 x 600, windowed.
14:17:58.707 [display.initGL]: Created a new surface of 800 x 600, windowed.
14:17:58.824 [display.initGL]: Requested a new surface of 1600 x 900, fullscreen.
14:17:58.829 [display.initGL]: Created a new surface of 1600 x 900, fullscreen.
14:17:58.957 [display.initGL]: Requested a new surface of 800 x 600, windowed.
14:17:58.989 [display.initGL]: Created a new surface of 800 x 600, windowed.
14:17:59.090 [display.initGL]: Requested a new surface of 1600 x 900, fullscreen.
14:17:59.095 [display.initGL]: Created a new surface of 1600 x 900, fullscreen.
14:17:59.240 [display.initGL]: Requested a new surface of 800 x 600, windowed.
14:17:59.273 [display.initGL]: Created a new surface of 800 x 600, windowed.
14:17:59.373 [display.initGL]: Requested a new surface of 1600 x 900, fullscreen.
14:17:59.378 [display.initGL]: Created a new surface of 1600 x 900, fullscreen.
14:17:59.540 [display.initGL]: Requested a new surface of 800 x 600, windowed.
14:17:59.574 [display.initGL]: Created a new surface of 800 x 600, windowed.
14:17:59.672 [display.initGL]: Requested a new surface of 1600 x 900, fullscreen.
14:17:59.677 [display.initGL]: Created a new surface of 1600 x 900, fullscreen.
14:17:59.806 [display.initGL]: Requested a new surface of 800 x 600, windowed.
14:17:59.839 [display.initGL]: Created a new surface of 800 x 600, windowed.
14:17:59.922 [display.initGL]: Requested a new surface of 1600 x 900, fullscreen.
14:17:59.927 [display.initGL]: Created a new surface of 1600 x 900, fullscreen.
14:18:00.089 [display.initGL]: Requested a new surface of 800 x 600, windowed.
14:18:00.123 [display.initGL]: Created a new surface of 800 x 600, windowed.
14:18:00.322 [display.initGL]: Requested a new surface of 1600 x 900, fullscreen.
14:18:00.326 [display.initGL]: Created a new surface of 1600 x 900, fullscreen.
14:18:11.258 [exit.context]: Exiting: Exit Game selected on start screen.
14:18:11.259 [gameController.exitApp]: .GNUstepDefaults synchronized.

Closing log at 2015-05-31 14:18:11 -0400.

EDIT #2: No access to the Ubuntu machine I initially replicated the case, so no uname/wmctrl info from there yet.

[jh145]

That's your second machine, isn't it? Are we just better decathletes than everyone else?!

I think we may have reached the end of the road with this. A big THANK YOU to everyone that tested their systems and posted config information, but sadly I can't see any sort of consistent pattern emerging. I don't know what to make of it at all.

[Getafix]

@jh145: Yes, I just made a Fedora 22 Live USB and tried it on a laptop (not the initial machine). I will also try Debian the same way.

[Getafix]

Could not replicate on Slax 7.0.8 and Debian 8.0.0 (edit to add: LXDE); both live USB on Lenovo T430 Intel HD4000.

Slax 7.0.8:

Code: Select all

root@slax:~# uname -rvp
3.8.2 #1 SMP Tue Mar 12 03:43:27 AKDT 2013 Intel(R) Core(TM) i5-3320M CPU @ 2.60GHz
root@slax:~# wmctrl -m
Name: KWin
Class: kwin
PID: 2352
Window manager's "showing the desktop" mode: OFF

Debian 8.0.0:

Code: Select all

user@debian:~$ uname -rvp
3.16.0-4-amd64 #1 SMP Debian 3.16.7-ckt9-2 (2015-04-13) unknown
user@debian:~$ wmctrl -m
Name: Openbox
Class:
PID: N/A
Window manager's "showing the desktop" mode: OFF

[jh145]

Thanks for those. Can you give me this machine as well please (fwiw):

I replicated the case on Ubuntu 14.04 64bit with nVidia 340.76 drivers GTX750Ti card.

[Commander_X]

Are you running unity?

I don't run unity (I run Slackware64, and unity is mainly available on Ubuntu).
The contents of your unity.ini file lists the compiz plugins loaded at startup, and that's not very useful from a plugin configuration standpoint.
I am mainly loading only the "compiz config plugin" (specified as ccp in compiz command line) and then I enable/disable/configure plugins using ccsm (compiz config settings manager). On my system the place where these settings are stored is

Code: Select all

$HOME/.config/compiz/compizconfig/Default.ini

and the full screen setting I've been thinking might help is

Code: Select all

s0_unredirect_fullscreen_windows = true

under the [core] section.

[Getafix]

Thanks for those. Can you give me this machine as well please (fwiw):

I replicated the case on Ubuntu 14.04 64bit with nVidia 340.76 drivers GTX750Ti card.

It's the default window manager of Ubuntu 14.04 Unity/Compiz/GNOME:

Code: Select all

getafix@Elite:~$ uname -rvp
3.16.0-30-generic #40~14.04.1-Ubuntu SMP Thu Jan 15 17:43:14 UTC 2015 x86_64
getafix@Elite:~$ wmctrl -m
Name: Compiz
Class: N/A
PID: N/A
Window manager's "showing the desktop" mode: OFF

There is something fishy around GNOME and Compiz.

[Getafix]

and the full screen setting I've been thinking might help is

Code: Select all

s0_unredirect_fullscreen_windows = true

...

Installed Compiz Settings Manager and this was set to TRUE by default.
Disabling Unity (yes...drastic measure) led to a multitude of plugins to be disabled, leaving a crippled bricked desktop. I still had a terminal window so I launched Oolite. Guess what... No F12 issue!

What a fine opportunity to install and try another distro!

[jh145]

and the full screen setting I've been thinking might help is

Code: Select all

s0_unredirect_fullscreen_windows = true

...

Installed Compiz Settings Manager and this was set to TRUE by default.
Disabling Unity (yes...drastic measure) led to a multitude of plugins to be disabled, leaving a crippled bricked desktop. I still had a terminal window so I launched Oolite. Guess what... No F12 issue!

Crippled bricked desktop?! You'll forgive me if I don't try to reproduce that.

Are you able to selectively disable plugins to narrow down the offending functionality? I'd do it myself, but I obviously know a lot less about window systems than you or Commander_X, so better that you forge ahead ...

[Getafix]

No no no... don't try this at home!
It was a vanilla installation that I didn't care much.
Initially I tried by disabling selectively some plugins but with no success.
Anyhow, I would say that it is compiz but then how did I manage to replicate it (multiple times) on the Fedora22 live USB?
Well, this is just me thinking loudly. I will check what did Fedora do with the Gnome they offer at their last release.

[Commander_X]

It seems you already had "crippled bricked desktops" already, due to a nice plugin which is part of the unity.ini file - i.e. unityshell at the end of it.
I'm not going to feed the internet's riot against Ubuntu/Unity, though, as I'm not even affected. But it seems some of the reasons were sound.

[Commander_X]

[...] but then how did I manage to replicate it (multiple times) on the Fedora22 live USB?
Well, this is just me thinking loudly. I will check what did Fedora do with the Gnome they offer at their last release.

I won't play the devil's advocate here, but Fedora is trying to push the other hated spawn of the gnome-desktop hell hole -- gnome-shell.
There is a reason these distributions have their Gnome2, KDE and Xfce variants.

[Getafix]

@jh145: One more buggy for the record. The O/S is LXLE 32bit on a Compaq Evo N800v.

Code: Select all

effie@sisyphus:~$ uname -rvp
3.2.0-84-generic #121-Ubuntu SMP Tue May 5 18:55:46 UTC 2015 i686
effie@sisyphus:~$ wmctrl -m
Name: Openbox
Class: 
PID: N/A
Window manager's "showing the desktop" mode: OFF

...I managed to find the same occurrence outside Oolite:
http://www.bay12games.com/dwarves/manti ... hp?id=7778 (at least here there is some info on a probable root-cause)
https://bbs.archlinux.org/viewtopic.php?id=128420 (no added value here...I give it just for your info)

The first link indicates that the issue involves SDL and XIM not being that much foolproof. This results in a small time-window where the user can request SDL to asynchronously change mode while the previous request is not completely served.

A sloppy workaround in Oolite, if this case is that much of a trouble, could be to try to reduce the frequency of firing up SDL by adding a delay. I am currently testing with various delays and it seems that a delay of 375msec makes it really difficult (not impossible though) to reproduce the case. Note, however, the the delay is perceivable... not annoying but perceivable.

I am talking about a workaround because it is a fact, for the time being, that the case cannot be reproduced in 1.80.

EDIT: Adding a delay is not a workaround. It fails to reduce the effect. Will try to find where was this inserted.

[jh145]

Great sleuthing!

I found another post implicating ibus. I'll see if I can make sense of that when I'm back on my otherwise-lovely Ubuntu box.

Update: I think I've got it! Just unset XMODIFIERS for Oolite:

Code: Select all

$ XMODIFIERS= ./oolite-wrapper
...

That probably has an adverse effect on accented or non-English character input to Oolite. Apart from choosing your commander name, I can't see how that would affect game-play.

For completeness, would you mind checking, Getafix, for the value of XMODIFIERS (for me: @im=ibus), the result of "ibus version" (for me: IBus 1.5.5) and the efficacy or otherwise of unsetting XMODIFIERS before running Oolite? Many thanks for all your help with this.

[Getafix]

@jh145: I will try this as soon as I get home.

In the meantime I think I have found the culprit for having that in Oolite (binary search rulez!).
It is called SDL_EnableUNICODE and we started using it the 5th of July... just four days after 1.80 release.
This added the possibility to enter non-alphanumeric characters in the text entry mission screens.

I will take it to The Elders Council to see if there are any scrolls suggesting something on the matter.