From thomas at koch.ro Tue Sep 3 21:25:45 2013 From: thomas at koch.ro (Thomas Koch) Date: Tue, 3 Sep 2013 22:25:45 +0200 Subject: [winswitch] is xpra secure? Message-ID: <201309032225.45426.thomas@koch.ro> Hi, I plan to use xpra to connect to untrusted software (Skype, Browser with Flash, Eclipse with plugins). I learned that SSH -x is not save for this purpose since clients could sniff my keyboard input or send arbitrary data to a terminal. Is xpra saver than SSH -x? Thank you, Thomas Koch From antoine at nagafix.co.uk Wed Sep 4 11:43:38 2013 From: antoine at nagafix.co.uk (Antoine Martin) Date: Wed, 04 Sep 2013 17:43:38 +0700 Subject: [winswitch] is xpra secure? In-Reply-To: <201309032225.45426.thomas@koch.ro> References: <201309032225.45426.thomas@koch.ro> Message-ID: <52270EDA.9050508@nagafix.co.uk> On 04/09/13 03:25, Thomas Koch wrote: > Hi, > > I plan to use xpra to connect to untrusted software (Skype, Browser with > Flash, Eclipse with plugins). I learned that SSH -x is not save for this > purpose since clients could sniff my keyboard input or send arbitrary data to a > terminal. > > Is xpra saver than SSH -x? Yes. Xpra only forwards the pixels, mouse and keyboard events are only forwarded whilst the window is in focus or directly under the mouse, the client applications do not have access to your local display at all. They are not even aware that it exists. Antoine From thomas at koch.ro Wed Sep 4 16:00:17 2013 From: thomas at koch.ro (Thomas Koch) Date: Wed, 4 Sep 2013 17:00:17 +0200 Subject: [winswitch] please conform to XDG Base directory specification; ~/.config/xpra/ Message-ID: <201309041700.17340.thomas@koch.ro> Hi, I'm continuing my exploration of xpra. It's wonderful so far. It would become even nicer, if you'd conform to the XDG Base Directory Specification[1][2]. That means the xpra config file would go to ~/.config/xpra/config [1] http://standards.freedesktop.org/basedir-spec/basedir-spec-latest.html [2] https://wiki.debian.org/XDGBaseDirectorySpecification Thank you! Thomas Koch From antoine at nagafix.co.uk Wed Sep 4 16:12:06 2013 From: antoine at nagafix.co.uk (Antoine Martin) Date: Wed, 04 Sep 2013 22:12:06 +0700 Subject: [winswitch] please conform to XDG Base directory specification; ~/.config/xpra/ In-Reply-To: <201309041700.17340.thomas@koch.ro> References: <201309041700.17340.thomas@koch.ro> Message-ID: <52274DC6.1030103@nagafix.co.uk> On 04/09/13 22:00, Thomas Koch wrote: > Hi, > > I'm continuing my exploration of xpra. It's wonderful so far. It would become > even nicer, if you'd conform to the XDG Base Directory Specification[1][2]. > That means the xpra config file would go to > > ~/.config/xpra/config I'm not very keen on this. OTOH I can think of two problems already: * we already have ~/.xpra/ shipping - changing anything is bound to cause compatibility problems. * the server sockets are placed in ~/.xpra/ by default already (and that is unlikely to change - and even if it did, .config is not a suitable place for sockets) Ideas welcome. Antoine > > [1] http://standards.freedesktop.org/basedir-spec/basedir-spec-latest.html > [2] https://wiki.debian.org/XDGBaseDirectorySpecification > > Thank you! > > Thomas Koch > _______________________________________________ > shifter-users mailing list > shifter-users at lists.devloop.org.uk > http://lists.devloop.org.uk/mailman/listinfo/shifter-users From thomas at koch.ro Wed Sep 4 16:39:24 2013 From: thomas at koch.ro (Thomas Koch) Date: Wed, 4 Sep 2013 17:39:24 +0200 Subject: [winswitch] please conform to XDG Base directory specification; ~/.config/xpra/ In-Reply-To: <52274DC6.1030103@nagafix.co.uk> References: <201309041700.17340.thomas@koch.ro> <52274DC6.1030103@nagafix.co.uk> Message-ID: <201309041739.24583.thomas@koch.ro> On Wednesday, September 04, 2013 05:12:06 PM Antoine Martin wrote: > On 04/09/13 22:00, Thomas Koch wrote: > > Hi, > > > > I'm continuing my exploration of xpra. It's wonderful so far. It would > > become even nicer, if you'd conform to the XDG Base Directory > > Specification[1][2]. That means the xpra config file would go to > > > > ~/.config/xpra/config > > I'm not very keen on this. > OTOH I can think of two problems already: > * we already have ~/.xpra/ shipping - changing anything is bound to > cause compatibility problems. > * the server sockets are placed in ~/.xpra/ by default already (and that > is unlikely to change - and even if it did, .config is not a suitable > place for sockets) > > Ideas welcome. xpra would not be the first application to move from ~/.something to XDG. Git is one example. The normal migration path is to have a fallback to also look for config files in the old location. The XDG spec also specifies the $XDG_RUNTIME_DIR for sockets, but unfortunately this directory is not yet accepted by distros AFAIK. But you could keep them in ~/.xpra until this is worked out. Thank you for your quick reply, Thomas Koch From thomas at koch.ro Thu Sep 5 08:28:04 2013 From: thomas at koch.ro (Thomas Koch) Date: Thu, 5 Sep 2013 09:28:04 +0200 Subject: [winswitch] connect retry loop for xpra launcher? Message-ID: <201309050928.04279.thomas@koch.ro> Hi, I'm working to setup xpra to connect to a virtual machine running on my laptop. The Host runs only trusted Debian stable software and the virtual machine runs all untrusted stuff (Skype...). Libvirtd automatically starts the virtual machine at boot and the xpra_launcher ist started by ~/.config/autostart/xpra_launcher.desktop However when xpra_launcher starts before the xpra server in the virtual machine, it opens a window. It would be nice to have an additional parameter for xpra_launcher like autoretry=$INTEGER If the parameter is set and not zero, then xpra_launcher will just sit in the systray and trying to reconnect every $INTEGER seconds. Regards, Thomas Koch From thomas at koch.ro Thu Sep 5 08:03:11 2013 From: thomas at koch.ro (Thomas Koch) Date: Thu, 05 Sep 2013 09:03:11 +0200 Subject: [winswitch] Bug#721898: please backport xpra Message-ID: <20130905070311.5538.78033.reportbug@koch.intern> Package: xpra Version: 0.10.1+dfsg-1 Severity: wishlist Hi, I've built the xpra package from the 0.10.1+dfsg-1 source package on Debian Wheezy and it works (almost) flawlessly. Only opengl is reported as unsupported on the server side (which is a KVM running SID + xpra from experimental). I attach the output from gl_check.py. This ticket (and the linked mail) seems related to the missing OpenGL support: http://xpra.org/trac/ticket/330 Thank you for working on xpra! Thomas Koch P.s. Thank you for gpg signing the git tags in the packaging repo! -- System Information: Debian Release: 7.1 APT prefers stable-updates APT policy: (500, 'stable-updates'), (500, 'stable') Architecture: amd64 (x86_64) Kernel: Linux 3.10-0.bpo.2-amd64 (SMP w/8 CPU cores) Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Versions of packages xpra depends on: ii libavcodec53 6:0.8.6-1 ii libc6 2.13-38 ii libgtk2.0-0 2.24.10-2 ii libswscale2 6:0.8.6-1 ii libvpx1 1.1.0-1 ii libx11-6 2:1.5.0-1+deb7u1 ii libx264-123 2:0.123.2189+git35cf912-1 ii libxcomposite1 1:0.4.3-2 ii libxdamage1 1:1.1.3-2 ii libxfixes3 1:5.0-4+deb7u1 ii libxrandr2 2:1.3.2-2+deb7u1 ii libxtst6 2:1.2.1-1+deb7u1 ii python 2.7.3-4 ii python-gtk2 2.24.0-3+b1 ii x11-xserver-utils 7.7~3 ii xserver-xorg-input-void 1:1.4.0-1+b1 ii xserver-xorg-video-dummy 1:0.3.5-2+b1 Versions of packages xpra recommends: ii openssh-client 1:6.0p1-4 ii python-imaging 1.1.7-4 pn python-webm ii ssh-askpass-gnome [ssh-askpass] 1:6.0p1-4 Versions of packages xpra suggests: ii gstreamer0.10-plugins-bad 0.10.23-7.1 ii gstreamer0.10-plugins-good 0.10.31-3+nmu1 ii gstreamer0.10-plugins-ugly 0.10.19-2+b2 ii openssh-server 1:6.0p1-4 ii pulseaudio 2.0-6.1 ii pulseaudio-utils 2.0-6.1 ii python-dbus 1.1.1-1 ii python-gst0.10 0.10.22-3 ii python-gtkglext1 1.1.0-9.1 -- no debconf information -------------- next part -------------- 2013-09-05 08:50:43,606 pygdkglext version=(1, 1, 0) 2013-09-05 08:50:43,607 pygdkglext OpenGL version=(1, 4) 2013-09-05 08:50:43,607 using display mode: ['SINGLE'] 2013-09-05 08:50:43,642 No OpenGL_accelerate module loaded: No module named OpenGL_accelerate 2013-09-05 08:50:43,681 Unable to load registered array format handler numeric: Traceback (most recent call last): File "/usr/lib/python2.7/dist-packages/OpenGL/arrays/formathandler.py", line 44, in loadPlugin plugin_class = entrypoint.load() File "/usr/lib/python2.7/dist-packages/OpenGL/plugins.py", line 14, in load return importByName( self.import_path ) File "/usr/lib/python2.7/dist-packages/OpenGL/plugins.py", line 28, in importByName module = __import__( ".".join(moduleName), {}, {}, moduleName) File "/usr/lib/python2.7/dist-packages/OpenGL/arrays/numeric.py", line 15, in raise ImportError( """No Numeric module present: %s"""%(err)) ImportError: No Numeric module present: No module named Numeric 2013-09-05 08:50:43,705 OpenGL Version: 3.0 Mesa 8.0.5 2013-09-05 08:50:43,720 GL Extension GL_ARB_shader_objects available 2013-09-05 08:50:43,726 found valid OpenGL version: 3.0 2013-09-05 08:50:43,726 OpenGL extensions found: GL_ARB_multisample, GL_EXT_abgr, GL_EXT_bgra, GL_EXT_blend_color, GL_EXT_blend_minmax, GL_EXT_blend_subtract, GL_EXT_copy_texture, GL_EXT_polygon_offset, GL_EXT_subtexture, GL_EXT_texture_object, GL_EXT_vertex_array, GL_EXT_compiled_vertex_array, GL_EXT_texture, GL_EXT_texture3D, GL_IBM_rasterpos_clip, GL_ARB_point_parameters, GL_EXT_draw_range_elements, GL_EXT_packed_pixels, GL_EXT_point_parameters, GL_EXT_rescale_normal, GL_EXT_separate_specular_color, GL_EXT_texture_edge_clamp, GL_SGIS_generate_mipmap, GL_SGIS_texture_border_clamp, GL_SGIS_texture_edge_clamp, GL_SGIS_texture_lod, GL_ARB_framebuffer_sRGB, GL_ARB_multitexture, GL_EXT_framebuffer_sRGB, GL_IBM_multimode_draw_arrays, GL_IBM_texture_mirrored_repeat, GL_3DFX_texture_compression_FXT1, GL_ARB_texture_cube_map, GL_ARB_texture_env_add, GL_ARB_transpose_matrix, GL_EXT_blend_func_separate, GL_EXT_fog_coord, GL_EXT_multi_draw_arrays, GL_EXT_secondary_color, GL_EXT_texture_env_add , GL_EXT_texture_filter_anisotropic, GL_EXT_texture_lod_bias, GL_INGR_blend_func_separate, GL_NV_blend_square, GL_NV_light_max_exponent, GL_NV_texgen_reflection, GL_NV_texture_env_combine4, GL_SUN_multi_draw_arrays, GL_ARB_texture_border_clamp, GL_ARB_texture_compression, GL_EXT_framebuffer_object, GL_EXT_texture_env_combine, GL_EXT_texture_env_dot3, GL_MESA_window_pos, GL_NV_packed_depth_stencil, GL_NV_texture_rectangle, GL_NV_vertex_program, GL_ARB_depth_texture, GL_ARB_occlusion_query, GL_ARB_shadow, GL_ARB_texture_env_combine, GL_ARB_texture_env_crossbar, GL_ARB_texture_env_dot3, GL_ARB_texture_mirrored_repeat, GL_ARB_window_pos, GL_ATI_envmap_bumpmap, GL_EXT_stencil_two_side, GL_EXT_texture_cube_map, GL_NV_depth_clamp, GL_NV_vertex_program1_1, GL_APPLE_packed_pixels, GL_APPLE_vertex_array_object, GL_ARB_draw_buffers, GL_ARB_fragment_program, GL_ARB_fragment_shader, GL_ARB_shader_objects, GL_ARB_vertex_program, GL_ARB_vertex_shader, GL_ATI_draw_buffers, GL_ATI_texture_env_combin e3, GL_ATI_texture_float, GL_EXT_shadow_funcs, GL_EXT_stencil_wrap, GL_MESA_pack_invert, GL_MESA_ycbcr_texture, GL_ARB_depth_clamp, GL_ARB_fragment_program_shadow, GL_ARB_half_float_pixel, GL_ARB_point_sprite, GL_ARB_shading_language_100, GL_ARB_sync, GL_ARB_texture_non_power_of_two, GL_ARB_vertex_buffer_object, GL_ATI_blend_equation_separate, GL_EXT_blend_equation_separate, GL_OES_read_format, GL_ARB_color_buffer_float, GL_ARB_pixel_buffer_object, GL_ARB_texture_compression_rgtc, GL_ARB_texture_float, GL_ARB_texture_rectangle, GL_EXT_packed_float, GL_EXT_pixel_buffer_object, GL_EXT_texture_compression_rgtc, GL_EXT_texture_rectangle, GL_EXT_texture_sRGB, GL_EXT_texture_shared_exponent, GL_ARB_framebuffer_object, GL_EXT_framebuffer_blit, GL_EXT_framebuffer_multisample, GL_EXT_packed_depth_stencil, GL_APPLE_object_purgeable, GL_ARB_vertex_array_object, GL_ATI_separate_stencil, GL_EXT_draw_buffers2, GL_EXT_gpu_program_parameters, GL_EXT_texture_array, GL_EXT_texture_integer, GL_EXT_tex ture_sRGB_decode, GL_EXT_timer_query, GL_OES_EGL_image, GL_MESA_texture_array, GL_ARB_copy_buffer, GL_ARB_depth_buffer_float, GL_ARB_half_float_vertex, GL_ARB_map_buffer_range, GL_ARB_texture_rg, GL_ARB_texture_swizzle, GL_ARB_vertex_array_bgra, GL_EXT_separate_shader_objects, GL_EXT_texture_swizzle, GL_EXT_vertex_array_bgra, GL_NV_conditional_render, GL_ARB_ES2_compatibility, GL_ARB_draw_elements_base_vertex, GL_ARB_explicit_attrib_location, GL_ARB_fragment_coord_conventions, GL_ARB_provoking_vertex, GL_ARB_sampler_objects, GL_ARB_seamless_cube_map, GL_ARB_shader_texture_lod, GL_EXT_provoking_vertex, GL_EXT_texture_snorm, GL_MESA_texture_signed_rgba, GL_ARB_robustness, GL_EXT_transform_feedback, 2013-09-05 08:50:43,726 vendor: Tungsten Graphics, Inc 2013-09-05 08:50:43,726 renderer: Mesa DRI Intel(R) Ivybridge Desktop 2013-09-05 08:50:43,726 shading language version: 1.30 2013-09-05 08:50:43,732 GLU extensions: GLU_EXT_nurbs_tessellator GLU_EXT_object_space_tess 2013-09-05 08:50:43,732 GLU version: 1.3 2013-09-05 08:50:43,732 All the required OpenGL functions are available: glActiveTexture, glTexSubImage2D, glTexCoord2i, glViewport, glMatrixMode, glLoadIdentity, glOrtho, glEnableClientState, glGenTextures, glDisable, glBindTexture, glPixelStorei, glEnable, glBegin, glFlush, glTexParameteri, glTexImage2D, glMultiTexCoord2i, glVertex2i, glEnd 2013-09-05 08:50:43,732 GL Extension GL_ARB_framebuffer_object available 2013-09-05 08:50:43,734 All the required OpenGL functions are available: GL_FRAMEBUFFER (36160), GL_COLOR_ATTACHMENT0 (36064), glGenFramebuffers, glBindFramebuffer, glFramebufferTexture2D 2013-09-05 08:50:43,734 Extension GL_ARB_texture_rectangle is present 2013-09-05 08:50:43,734 Extension GL_ARB_vertex_program is present 2013-09-05 08:50:43,734 GL Extension GL_ARB_fragment_program available 2013-09-05 08:50:43,734 glInitFragmentProgramARB works 2013-09-05 08:50:43,734 GL Extension GL_ARB_texture_rectangle available 2013-09-05 08:50:43,734 glInitTextureRectangleARB works 2013-09-05 08:50:43,734 GL Extension GL_ARB_vertex_program available 2013-09-05 08:50:43,737 All the required OpenGL functions are available: glGenProgramsARB, glDeleteProgramsARB, glBindProgramARB, glProgramStringARB 2013-09-05 08:50:43,737 Texture size GL_MAX_RECTANGLE_TEXTURE_SIZE_ARB=4096 From antoine at nagafix.co.uk Fri Sep 6 05:45:21 2013 From: antoine at nagafix.co.uk (Antoine Martin) Date: Fri, 06 Sep 2013 11:45:21 +0700 Subject: [winswitch] [ANNOUNCE] xpra 0.10.3 (minor bug fixes) Message-ID: <52295DE1.5040306@nagafix.co.uk> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, This fixes only two fairly minor bugs: * fix transient windows with no parent (which were not showing) * fix metadata updates handling (snapping out of maximize, etc) The source: https://xpra.org/src/ Binaries/repositories: https://winswitch.org/downloads/ Direct binary downloads: https://xpra.org/dists/ Cheers Antoine -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.14 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iEYEARECAAYFAlIpXd0ACgkQGK2zHPGK1rt0IgCfU9s7aZ5m/xe+aSU59oK7Fg13 XbcAnjJsbguWJi9GLP7fBnz6sBHFVEtk =5AVR -----END PGP SIGNATURE----- From antoine at nagafix.co.uk Fri Sep 6 07:19:03 2013 From: antoine at nagafix.co.uk (Antoine Martin) Date: Fri, 06 Sep 2013 13:19:03 +0700 Subject: [winswitch] connect retry loop for xpra launcher? In-Reply-To: <201309050928.04279.thomas@koch.ro> References: <201309050928.04279.thomas@koch.ro> Message-ID: <522973D7.4000705@nagafix.co.uk> On 05/09/13 14:28, Thomas Koch wrote: > Hi, > > I'm working to setup xpra to connect to a virtual machine running on my > laptop. The Host runs only trusted Debian stable software and the virtual > machine runs all untrusted stuff (Skype...). > > Libvirtd automatically starts the virtual machine at boot and the > xpra_launcher ist started by ~/.config/autostart/xpra_launcher.desktop How so? Via the GUI? Autostart? > However when xpra_launcher starts before the xpra server in the virtual > machine, it opens a window. I don't understand why the launcher would *not* show the GUI unless you instruct it to open a session file. But you're not telling us anything about that. > It would be nice to have an additional parameter for xpra_launcher like > autoretry=$INTEGER Sounds to me like you just need to fire the connection when the VM has started. This is probably easier to implement than retries. The retries could more easily be implement as a wrapper script around /usr/bin/xpra > If the parameter is set and not zero, then xpra_launcher will just sit in the > systray and trying to reconnect every $INTEGER seconds. Feel free to file feature request ticket, but I don't think there is much interest in this feature at present. Cheers Antoine > > Regards, > > Thomas Koch > _______________________________________________ > shifter-users mailing list > shifter-users at lists.devloop.org.uk > http://lists.devloop.org.uk/mailman/listinfo/shifter-users From antoine at nagafix.co.uk Fri Sep 6 07:29:08 2013 From: antoine at nagafix.co.uk (Antoine Martin) Date: Fri, 06 Sep 2013 13:29:08 +0700 Subject: [winswitch] Bug#721898: please backport xpra In-Reply-To: <20130905070311.5538.78033.reportbug@koch.intern> References: <20130905070311.5538.78033.reportbug@koch.intern> Message-ID: <52297634.5010507@nagafix.co.uk> > Only opengl is reported as > unsupported on the server side (which is a KVM running SID + xpra from > experimental). Server side does not use OpenGL, maybe you are referring to this client side message: ImportError: No Numeric module present: No module named Numeric Note: Debian does not ship python-gnumeric (which is superseded by numpy) which makes PyOpenGL spit out a warning when we load it. AFAIK, silencing this warning would silence other more useful warnings. Perhaps Debian should consider patching PyOpenGL to remove this warning instead - or provide python-numeric packages for backwards compatibility. > I attach the output from gl_check.py. The output looks fine, and OpenGL should have been enabled OK on this system. You can check in your tray menu or in the Session-Info's Statistics window. > This ticket (and the linked mail) seems > related to the missing OpenGL support: http://xpra.org/trac/ticket/330 The ticket you are pointing to has nothing to do with GL... Antoine From onlyjob at member.fsf.org Fri Sep 6 10:54:56 2013 From: onlyjob at member.fsf.org (Dmitry Smirnov) Date: Fri, 6 Sep 2013 19:54:56 +1000 Subject: [winswitch] Bug#721898: please backport xpra In-Reply-To: <52297634.5010507@nagafix.co.uk> References: <20130905070311.5538.78033.reportbug@koch.intern> <52297634.5010507@nagafix.co.uk> Message-ID: <201309061954.59681.onlyjob@member.fsf.org> On Fri, 6 Sep 2013 16:29:08 Antoine Martin wrote: > Server side does not use OpenGL, maybe you are referring to this client > side message: > > ImportError: No Numeric module present: No module named Numeric > > > Note: Debian does not ship python-gnumeric (which is superseded by > numpy) which makes PyOpenGL spit out a warning when we load it. > AFAIK, silencing this warning would silence other more useful warnings. I wonder if logging is unnecessary verbose by default. For example the following command python -c 'import logging; logging.root.setLevel(logging.DEBUG); import OpenGL; from OpenGL.GL import GL_VERSION, GL_EXTENSIONS;' merely print No handlers could be found for logger "OpenGL.acceleratesupport" and somehow don't complain about "No module named Numeric"... > Perhaps Debian should consider patching PyOpenGL to remove this warning > instead - or provide python-numeric packages for backwards compatibility. It is up to "pyopengl" maintainer(s) to decide and they are already have a bug for this: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=703585 -- Best wishes, Dmitry Smirnov GPG key : 4096R/53968D1B From antoine at nagafix.co.uk Fri Sep 6 11:07:13 2013 From: antoine at nagafix.co.uk (Antoine Martin) Date: Fri, 06 Sep 2013 17:07:13 +0700 Subject: [winswitch] Bug#721898: Bug#721898: please backport xpra In-Reply-To: <201309061954.59681.onlyjob@member.fsf.org> References: <20130905070311.5538.78033.reportbug@koch.intern> <52297634.5010507@nagafix.co.uk> <201309061954.59681.onlyjob@member.fsf.org> Message-ID: <5229A951.6040507@nagafix.co.uk> On 06/09/13 16:54, Dmitry Smirnov wrote: > On Fri, 6 Sep 2013 16:29:08 Antoine Martin wrote: >> Server side does not use OpenGL, maybe you are referring to this client >> side message: >> >> ImportError: No Numeric module present: No module named Numeric >> >> >> Note: Debian does not ship python-gnumeric (which is superseded by >> numpy) which makes PyOpenGL spit out a warning when we load it. >> AFAIK, silencing this warning would silence other more useful warnings. > I wonder if logging is unnecessary verbose by default. For example the > following command > > python -c 'import logging; logging.root.setLevel(logging.DEBUG); import OpenGL; from OpenGL.GL import GL_VERSION, GL_EXTENSIONS;' > > merely print > > No handlers could be found for logger "OpenGL.acceleratesupport" > > and somehow don't complain about "No module named Numeric"... > That's a different message caused by the fact that the logging system is not configured before use. You need a call to "logging.basicConfig() or similar before importing OpenGL: python -c 'import logging; logging.basicConfig(); logging.root.setLevel(logging.DEBUG); import OpenGL; from OpenGL.GL import GL_VERSION, GL_EXTENSIONS;' Antoine From thomas at koch.ro Sat Sep 7 09:14:21 2013 From: thomas at koch.ro (Thomas Koch) Date: Sat, 7 Sep 2013 10:14:21 +0200 Subject: [winswitch] very high CPU use of xpra Message-ID: <201309071014.21456.thomas@koch.ro> Hi, xpra uses a lot of CPU resources on my virtual server, running the xpra debian package 0.10.1+dfsg-1 rebuilt against Wheezy stable. Xpra spawns 10 child threads (processes?) when one program (like firefox) is running. Three of the child threads seem to run a loop at full steam and consume each around 10% of CPU resources (see below). Also the pulseaudio process consumes around 10%. Is there anything I can do to further investigate this or are you aware of this problem. 1. Thread: clock_gettime(CLOCK_MONOTONIC, {1696, 356227712}) = 0 futex(0x34a4870, FUTEX_WAKE_PRIVATE, 1) = 1 futex(0x2a103b0, FUTEX_WAKE_PRIVATE, 1) = 1 futex(0x7f0ca406b3c4, FUTEX_WAIT_PRIVATE, 9795, NULL) = 0 futex(0x7f0ca4085650, FUTEX_WAKE_PRIVATE, 1) = 0 futex(0x7f0ca406b3c4, FUTEX_WAIT_PRIVATE, 9797, NULL) = 0 futex(0x7f0ca4085650, FUTEX_WAKE_PRIVATE, 1) = 0 clock_gettime(CLOCK_MONOTONIC, {1696, 375617815}) = 0 2. Thread munmap(0x7f0cabfff000, 67112960) = 0 write(20, "W", 1) = 1 futex(0x34035f0, FUTEX_UNLOCK_PI_PRIVATE, 33) = 0 futex(0x34b3de4, FUTEX_WAIT_REQUEUE_PI_PRIVATE, 88035, NULL, 0x34035f0) = 0 munmap(0x7f0cabfff000, 67112960) = 0 write(20, "W", 1) = 1 futex(0x34035f0, FUTEX_UNLOCK_PI_PRIVATE, 33) = 0 futex(0x7f0ca406b3c4, FUTEX_WAKE_OP_PRIVATE, 1, 1, 0x7f0ca406b3c0, {FUTEX_OP_SET, 0, FUTEX_OP_CMP_GT, 1}) = 1 futex(0x7f0ca4085650, FUTEX_WAKE_PRIVATE, 1) = 1 futex(0x34b3de4, FUTEX_WAIT_REQUEUE_PI_PRIVATE, 88037, NULL, 0x34035f0) = -1 EAGAIN (Resource temporarily unavailable) futex(0x34b3de4, FUTEX_WAIT_PRIVATE, 88037, NULL) = -1 EAGAIN (Resource temporarily unavailable) write(20, "W", 1) = 1 munmap(0x7f0cabfff000, 67112960) = 0 futex(0x34035f0, FUTEX_UNLOCK_PI_PRIVATE, 33) = 0 futex(0x7f0ca406b3c4, FUTEX_WAKE_OP_PRIVATE, 1, 1, 0x7f0ca406b3c0, {FUTEX_OP_SET, 0, FUTEX_OP_CMP_GT, 1}) = 1 3. Thread sendto(24, "\0\0\0\0\377\377\377\377\0\0\0\20\0\0\0\0@\0\0\0", 20, MSG_NOSIGNAL, NULL, 0) = 20 read(19, "W", 10) = 1 clock_gettime(CLOCK_MONOTONIC, {1908, 665097131}) = 0 poll([{fd=19, events=POLLIN}, {fd=24, events=POLLIN|POLLOUT}], 2, 1245) = 1 ([{fd=24, revents=POLLIN|POLLOUT}]) clock_gettime(CLOCK_MONOTONIC, {1908, 665496657}) = 0 recvmsg(24, {msg_name(0)=NULL, msg_iov(1)=[{"\0\0\0\20\0\0\0\0\0\0\0\0\0\0\0\0\200\0\0\0", 20}], msg_controllen=32, {cmsg_len=28, cmsg clock_gettime(CLOCK_MONOTONIC, {1908, 665652803}) = 0 poll([{fd=19, events=POLLIN}, {fd=24, events=POLLIN}], 2, 1244) = 1 ([{fd=24, revents=POLLIN}]) clock_gettime(CLOCK_MONOTONIC, {1908, 665751495}) = 0 recvmsg(24, {msg_name(0)=NULL, msg_iov(1)=[{"\0\0\0\17\366\357\tz\0\0\0@\0\0\3p", 16}], msg_controllen=32, {cmsg_len=28, cmsg_level=SO open("/dev/shm/pulse-shm-4142860666", O_RDONLY|O_NOFOLLOW|O_CLOEXEC) = 25 fstat(25, {st_mode=S_IFREG|0400, st_size=67108904, ...}) = 0 mmap(NULL, 67112960, PROT_READ, MAP_SHARED, 25, 0) = 0x7f0cabfff000 close(25) = 0 futex(0x34b3de4, FUTEX_CMP_REQUEUE_PI_PRIVATE, 1, 2147483647, 0x34035f0, 100926) = 1 clock_gettime(CLOCK_MONOTONIC, {1908, 666261504}) = 0 futex(0x34035f0, FUTEX_UNLOCK_PI_PRIVATE, 33) = 0 poll([{fd=19, events=POLLIN}, {fd=24, events=POLLIN}], 2, 1244) = 1 ([{fd=19, revents=POLLIN}]) sendto(24, "\0\0\0\0\377\377\377\377\0\0\0\17\0\0\0\0@\0\0\0", 20, MSG_NOSIGNAL, NULL, 0) = 20 read(19, "W", 10) = 1 clock_gettime(CLOCK_MONOTONIC, {1908, 666757989}) = 0 4. Pulseaudio Thread read(14, "\1\0\0\0\0\0\0\0", 8) = 8 sendto(18, "\0\0\0\20\0\0\0\0\0\0\0\0\0\0\0\0\200\0\0\0", 20, MSG_NOSIGNAL, NULL, 0) = 20 ppoll([{fd=7, events=POLLIN}, {fd=18, events=POLLIN|POLLOUT}, {fd=17, events=POLLIN}, {fd=16, events=POLLIN|POLLERR|POLLHUP}, {fd=16, sendto(18, "\0\0\0\6\366\357\tz\0\0\0@\0\0\3p", 16, MSG_NOSIGNAL, NULL, 0) = 16 ppoll([{fd=7, events=POLLIN}, {fd=18, events=POLLIN|POLLOUT}, {fd=17, events=POLLIN}, {fd=16, events=POLLIN|POLLERR|POLLHUP}, {fd=16, recvmsg(18, {msg_name(0)=NULL, msg_iov(1)=[{"\0\0\0\0\377\377\377\377\0\0\0\6\0\0\0\0@\0\0\0", 20}], msg_controllen=32, {cmsg_len=28, read(14, "\1\0\0\0\0\0\0\0", 8) = 8 sendto(18, "\0\0\0\20\0\0\0\0\0\0\0\0\0\0\0\0\200\0\0\0", 20, MSG_NOSIGNAL, NULL, 0) = 20 Regards, Thomas Koch From thomas at koch.ro Sat Sep 7 14:07:26 2013 From: thomas at koch.ro (Thomas Koch) Date: Sat, 7 Sep 2013 15:07:26 +0200 Subject: [winswitch] Use spice to show individual X clients like xpra does? Message-ID: <201309071507.26562.thomas@koch.ro> Hi, I've played around a bit with xpra[1]. Xpra (AFAIK) starts an X server with a dummy video driver (or xvfb?) on the remote machine and a client xpra process on the local machine shows remote application windows just like they would run locally. [1] http://xpra.org Xpra also forwards sound, clipboard and microfon. So it seems to me that there is some overlap with spice. It would be cool, if I could use spice in the same way, to see virtual machine X clients integrated in my local desktop just like local applications. Is this on your roadmap? Regards, Thomas Koch From antoine at nagafix.co.uk Sun Sep 8 13:06:53 2013 From: antoine at nagafix.co.uk (Antoine Martin) Date: Sun, 08 Sep 2013 19:06:53 +0700 Subject: [winswitch] very high CPU use of xpra In-Reply-To: <201309071014.21456.thomas@koch.ro> References: <201309071014.21456.thomas@koch.ro> Message-ID: <522C685D.3080703@nagafix.co.uk> On 07/09/13 15:14, Thomas Koch wrote: > Hi, > > xpra uses a lot of CPU resources on my virtual server, running the xpra debian package 0.10.1+dfsg-1 rebuilt > against Wheezy stable. Xpra spawns 10 child threads (processes?) when one program (like firefox) is running. The Xpra *server* spawns 6+ threads on top of the main thread: * 4 for networking, 2 in and 2 out: one for IO and one for (de)compression/parsing or formatting - *for each client connected*. * one for encoding pixel data (x264, png, etc) * one for background performance calculations (only if encoding is in use) * gstreamer will create a few threads of its own to handle sound (if enabled) Since the main thread handles the UI, it is therefore active at least when the UI is (and more), all the other threads operate as workqueues and will not be active unless they have something to do - that is unless there is something to compress or send/receive, etc. What is less obvious is that even when there is no sound being played, we still get sound data from pulseaudio and route it compressed to the client, which also uses up resources (again, only if enabled). > Three of the child threads seem to run a loop at full steam and consume each around 10% of CPU resources > (see below). "loop at full steam [...] 10% of CPU" - those two statements contradict each other. > Also the pulseaudio process consumes around 10%. Then turn sound off? Or complain to pulseaudio? Not much *we* can do. > Is there anything I can do to further > investigate this or are you aware of this problem. Those are probably the sound threads. If so, turn off sound and they will disappear completely. Note: firefox also has some little widgets that are constantly spinning and causing screen updates, it could be that too or at least making it worse. If the window is not shown (ie: minimized) then those screen updates will also stop causing CPU and network load since they will not be forwarded to the client. > (snip) Cheers Antoine From timo.lindfors at iki.fi Sun Sep 8 16:34:20 2013 From: timo.lindfors at iki.fi (Timo Juhani Lindfors) Date: Sun, 08 Sep 2013 18:34:20 +0300 Subject: [winswitch] is xpra secure? In-Reply-To: <52270EDA.9050508@nagafix.co.uk> (Antoine Martin's message of "Wed, 04 Sep 2013 17:43:38 +0700") References: <201309032225.45426.thomas@koch.ro> <52270EDA.9050508@nagafix.co.uk> Message-ID: <84li3749ab.fsf@sauna.l.org> Antoine Martin writes: > Xpra only forwards the pixels, mouse and keyboard events are only > forwarded whilst the window is in focus or directly under the mouse, the > client applications do not have access to your local display at all. > They are not even aware that it exists. It also forwards freedesktop.org notifications. From antoine at nagafix.co.uk Sun Sep 8 16:51:52 2013 From: antoine at nagafix.co.uk (Antoine Martin) Date: Sun, 08 Sep 2013 22:51:52 +0700 Subject: [winswitch] is xpra secure? In-Reply-To: <84li3749ab.fsf@sauna.l.org> References: <201309032225.45426.thomas@koch.ro> <52270EDA.9050508@nagafix.co.uk> <84li3749ab.fsf@sauna.l.org> Message-ID: <522C9D18.2010701@nagafix.co.uk> On 08/09/13 22:34, Timo Juhani Lindfors wrote: > Antoine Martin writes: >> Xpra only forwards the pixels, mouse and keyboard events are only >> forwarded whilst the window is in focus or directly under the mouse, the >> client applications do not have access to your local display at all. >> They are not even aware that it exists. > It also forwards freedesktop.org notifications. It also forwards bell events, sound, cursors and many other things from applications to the client. But these are all irrelevant in the context of the OP's question: none of this allows any application access to the client's display. The only data that travels from the client to the application is what I listed above. Antoine From mlureau at redhat.com Mon Sep 9 18:42:00 2013 From: mlureau at redhat.com (=?utf-8?Q?Marc-Andr=C3=A9?= Lureau) Date: Mon, 9 Sep 2013 13:42:00 -0400 (EDT) Subject: [winswitch] [Spice-devel] Use spice to show individual X clients like xpra does? In-Reply-To: <201309071507.26562.thomas@koch.ro> References: <201309071507.26562.thomas@koch.ro> Message-ID: <644027171.9898104.1378748520392.JavaMail.root@redhat.com> Hi ----- Original Message ----- > Hi, > > I've played around a bit with xpra[1]. Xpra (AFAIK) starts an X server with a > dummy video driver (or xvfb?) on the remote machine and a client xpra process > on the local machine shows remote application windows just like they would > run > locally. > > [1] http://xpra.org > > Xpra also forwards sound, clipboard and microfon. So it seems to me that > there > is some overlap with spice. > > It would be cool, if I could use spice in the same way, to see virtual > machine > X clients integrated in my local desktop just like local applications. Is > this > on your roadmap? It would be really nice to have, but nobody is working on this afaik. The current spice project is more designed toward driver level, rather than application level. You'd need to have a per-window drawing context in order to be able to display the seamless apps correctly. On Windows, it would probably need user-space gdi (&directx) hooks (somehow a bit like wine). Older methods just clip the remote windows with help of the window manager, however this is quite limited (see seamlessrdp or virtualbox seamless mode etc). I haven't looked at xpra, it would be interesting to see if and how they manage this per-window context. Perhaps it is possible to extend Xspice similarly? but I suppose it is rather done at higher level (application or X server frontend). From alevy at redhat.com Mon Sep 9 19:16:52 2013 From: alevy at redhat.com (Alon Levy) Date: Mon, 9 Sep 2013 14:16:52 -0400 (EDT) Subject: [winswitch] [Spice-devel] Use spice to show individual X clients like xpra does? In-Reply-To: <644027171.9898104.1378748520392.JavaMail.root@redhat.com> References: <201309071507.26562.thomas@koch.ro> <644027171.9898104.1378748520392.JavaMail.root@redhat.com> Message-ID: <1042921643.10293294.1378750612273.JavaMail.root@redhat.com> > > Hi > > ----- Original Message ----- > > Hi, > > > > I've played around a bit with xpra[1]. Xpra (AFAIK) starts an X server with > > a > > dummy video driver (or xvfb?) on the remote machine and a client xpra > > process > > on the local machine shows remote application windows just like they would > > run > > locally. > > > > [1] http://xpra.org > > > > Xpra also forwards sound, clipboard and microfon. So it seems to me that > > there > > is some overlap with spice. > > > > It would be cool, if I could use spice in the same way, to see virtual > > machine > > X clients integrated in my local desktop just like local applications. Is > > this > > on your roadmap? > > It would be really nice to have, but nobody is working on this afaik. > > The current spice project is more designed toward driver level, rather than > application level. You'd need to have a per-window drawing context in order > to be able to display the seamless apps correctly. On Windows, it would > probably need user-space gdi (&directx) hooks (somehow a bit like wine). > Older methods just clip the remote windows with help of the window manager, > however this is quite limited (see seamlessrdp or virtualbox seamless mode > etc). I haven't looked at xpra, it would be interesting to see if and how > they manage this per-window context. Perhaps it is possible to extend Xspice > similarly? but I suppose it is rather done at higher level (application or X > server frontend). See http://xpra.org/trac/browser/xpra/trunk/src/README I have to confess I used it some time in the past and completely forgot about it, thanks for bringing it up! It is of course X specific, it uses Xvfb and becomes a compositing manager on top of it, and instead of compositing forwards the images to the remote X server, and also acts as a window manager deferring all app requests to the remote window manager on the client X. I haven't looked at the protocol, or copy paste support, or audio, which at this point Xspice supports. It should be easy to use the compositing manager + window manager bits (not sure about language). And teach the client of course to implement them too. Any takers? > _______________________________________________ > Spice-devel mailing list > Spice-devel at lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/spice-devel > From antoine at nagafix.co.uk Tue Sep 10 08:22:13 2013 From: antoine at nagafix.co.uk (Antoine Martin) Date: Tue, 10 Sep 2013 14:22:13 +0700 Subject: [winswitch] [ANNOUNCE] xpra 0.10.4 (minor bug fixes) Message-ID: <522EC8A5.8080406@nagafix.co.uk> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, This fixes only two minor bugs: * fix modifier key handling (more noticeable with MS Windows clients) * fix auto-refresh The source: https://xpra.org/src/ Binaries/repositories: https://winswitch.org/downloads/ Direct binary downloads: https://xpra.org/dists/ Cheers Antoine -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.14 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iEYEARECAAYFAlIuyKIACgkQGK2zHPGK1rsiYQCeKy8NWEmDoPm5KbMQTlnD8Go1 Vt4An2j7Uk36vPBCGI2PIGOAUaQEy20t =INUo -----END PGP SIGNATURE----- From antoine at nagafix.co.uk Tue Sep 10 15:31:10 2013 From: antoine at nagafix.co.uk (Antoine Martin) Date: Tue, 10 Sep 2013 21:31:10 +0700 Subject: [winswitch] [Spice-devel] Use spice to show individual X clients like xpra does? In-Reply-To: <1042921643.10293294.1378750612273.JavaMail.root@redhat.com> References: <201309071507.26562.thomas@koch.ro> <644027171.9898104.1378748520392.JavaMail.root@redhat.com> <1042921643.10293294.1378750612273.JavaMail.root@redhat.com> Message-ID: <522F2D2E.8030206@nagafix.co.uk> On 10/09/13 01:16, Alon Levy wrote: >> Hi >> >> ----- Original Message ----- >>> Hi, >>> >>> I've played around a bit with xpra[1]. Xpra (AFAIK) starts an X server with >>> a >>> dummy video driver (or xvfb?) on the remote machine and a client xpra >>> process >>> on the local machine shows remote application windows just like they would >>> run >>> locally. >>> >>> [1] http://xpra.org >>> >>> Xpra also forwards sound, clipboard and microfon. So it seems to me that >>> there >>> is some overlap with spice. >>> >>> It would be cool, if I could use spice in the same way, to see virtual >>> machine >>> X clients integrated in my local desktop just like local applications. Is >>> this >>> on your roadmap? >> It would be really nice to have, but nobody is working on this afaik. >> >> The current spice project is more designed toward driver level, rather than >> application level. You'd need to have a per-window drawing context in order >> to be able to display the seamless apps correctly. On Windows, it would >> probably need user-space gdi (&directx) hooks (somehow a bit like wine). >> Older methods just clip the remote windows with help of the window manager, >> however this is quite limited (see seamlessrdp or virtualbox seamless mode >> etc). I haven't looked at xpra, it would be interesting to see if and how >> they manage this per-window context. Perhaps it is possible to extend Xspice >> similarly? but I suppose it is rather done at higher level (application or X >> server frontend). > See http://xpra.org/trac/browser/xpra/trunk/src/README > I have to confess I used it some time in the past and completely forgot about it, thanks for bringing it up! > It is of course X specific, Mostly yes, though the codebase no longer has any hard X11 dependencies for servers: the shadow mode works with MS Windows servers too for example (though at present it is no more than a cheap x0vncserver equivalent - just screen scraping). (and there are clients in python and Java for many platforms) > it uses Xvfb and becomes a compositing manager on top of it, and instead of compositing forwards the images to the remote X server, and also acts as a window manager deferring all app requests to the remote window manager on the client X. I haven't looked at the protocol, The wire protocol is pretty simple, and supports bencoded data (or rencoded data which is much faster) with a very limited set of commands for mapping/moving windows, etc. I guess another feature that would be relevant to you would be notification forwarding, the rest is probably handled already. > or copy paste support, Is fully supported on all client platforms, but is one way only for OSX clients at present. > or audio, We use gstreamer to compress the audio, usually as mp3 via lame. > which at this point Xspice supports. It should be easy to use the compositing manager + window manager bits (not sure about language). "Easy" is not the word I would have chosen here ;) But maybe spice makes it easier somehow? Antoine > And teach the client of course to implement them too. Any takers? > >> _______________________________________________ >> Spice-devel mailing list >> Spice-devel at lists.freedesktop.org >> http://lists.freedesktop.org/mailman/listinfo/spice-devel >> > _______________________________________________ > shifter-users mailing list > shifter-users at lists.devloop.org.uk > http://lists.devloop.org.uk/mailman/listinfo/shifter-users From alevy at redhat.com Tue Sep 10 17:21:32 2013 From: alevy at redhat.com (Alon Levy) Date: Tue, 10 Sep 2013 12:21:32 -0400 (EDT) Subject: [winswitch] [Spice-devel] Use spice to show individual X clients like xpra does? In-Reply-To: <522F2D2E.8030206@nagafix.co.uk> References: <201309071507.26562.thomas@koch.ro> <644027171.9898104.1378748520392.JavaMail.root@redhat.com> <1042921643.10293294.1378750612273.JavaMail.root@redhat.com> <522F2D2E.8030206@nagafix.co.uk> Message-ID: <1723206949.10922890.1378830092854.JavaMail.root@redhat.com> > On 10/09/13 01:16, Alon Levy wrote: > >> Hi > >> > >> ----- Original Message ----- > >>> Hi, > >>> > >>> I've played around a bit with xpra[1]. Xpra (AFAIK) starts an X server > >>> with > >>> a > >>> dummy video driver (or xvfb?) on the remote machine and a client xpra > >>> process > >>> on the local machine shows remote application windows just like they > >>> would > >>> run > >>> locally. > >>> > >>> [1] http://xpra.org > >>> > >>> Xpra also forwards sound, clipboard and microfon. So it seems to me that > >>> there > >>> is some overlap with spice. > >>> > >>> It would be cool, if I could use spice in the same way, to see virtual > >>> machine > >>> X clients integrated in my local desktop just like local applications. Is > >>> this > >>> on your roadmap? > >> It would be really nice to have, but nobody is working on this afaik. > >> > >> The current spice project is more designed toward driver level, rather > >> than > >> application level. You'd need to have a per-window drawing context in > >> order > >> to be able to display the seamless apps correctly. On Windows, it would > >> probably need user-space gdi (&directx) hooks (somehow a bit like wine). > >> Older methods just clip the remote windows with help of the window > >> manager, > >> however this is quite limited (see seamlessrdp or virtualbox seamless mode > >> etc). I haven't looked at xpra, it would be interesting to see if and how > >> they manage this per-window context. Perhaps it is possible to extend > >> Xspice > >> similarly? but I suppose it is rather done at higher level (application or > >> X > >> server frontend). > > See http://xpra.org/trac/browser/xpra/trunk/src/README > > I have to confess I used it some time in the past and completely forgot > > about it, thanks for bringing it up! > > It is of course X specific, > Mostly yes, though the codebase no longer has any hard X11 dependencies > for servers: the shadow mode works with MS Windows servers too for > example (though at present it is no more than a cheap x0vncserver > equivalent - just screen scraping). > (and there are clients in python and Java for many platforms) > > it uses Xvfb and becomes a compositing manager on top of it, and instead > > of compositing forwards the images to the remote X server, and also acts > > as a window manager deferring all app requests to the remote window > > manager on the client X. I haven't looked at the protocol, > The wire protocol is pretty simple, and supports bencoded data (or > rencoded data which is much faster) with a very limited set of commands > for mapping/moving windows, etc. > I guess another feature that would be relevant to you would be > notification forwarding, the rest is probably handled already. > > or copy paste support, > Is fully supported on all client platforms, but is one way only for OSX > clients at present. > > or audio, > We use gstreamer to compress the audio, usually as mp3 via lame. > > which at this point Xspice supports. It should be easy to use the > > compositing manager + window manager bits (not sure about language). > "Easy" is not the word I would have chosen here ;) > But maybe spice makes it easier somehow? You got me, I was using "easy" as in "easy to show but there isn't enough room in this email to include the proof". I mostly assumed I could take it as is, and would be left with introducing the protocol bits. > > Antoine > > And teach the client of course to implement them too. Any takers? > > > >> _______________________________________________ > >> Spice-devel mailing list > >> Spice-devel at lists.freedesktop.org > >> http://lists.freedesktop.org/mailman/listinfo/spice-devel > >> > > _______________________________________________ > > shifter-users mailing list > > shifter-users at lists.devloop.org.uk > > http://lists.devloop.org.uk/mailman/listinfo/shifter-users > > _______________________________________________ > Spice-devel mailing list > Spice-devel at lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/spice-devel > From thomas at koch.ro Sun Sep 15 17:07:47 2013 From: thomas at koch.ro (Thomas Koch) Date: Sun, 15 Sep 2013 18:07:47 +0200 Subject: [winswitch] does xpra raise windows? Message-ID: <201309151807.47923.thomas@koch.ro> Is xpra supposed to support raising X clients? If this is not implemented yet, then it would be a very nice feature to have. Thank you, Thomas Koch From antoine at nagafix.co.uk Mon Sep 16 06:19:36 2013 From: antoine at nagafix.co.uk (Antoine Martin) Date: Mon, 16 Sep 2013 12:19:36 +0700 Subject: [winswitch] does xpra raise windows? In-Reply-To: <201309151807.47923.thomas@koch.ro> References: <201309151807.47923.thomas@koch.ro> Message-ID: <523694E8.8050108@nagafix.co.uk> On 15/09/13 23:07, Thomas Koch wrote: > Is xpra supposed to support raising X clients? If this is not implemented yet, > then it would be a very nice feature to have. You're right, it wasn't implemented, it is now: http://xpra.org/trac/changeset/4335/xpra Antoine From chris at ambigc.com Mon Sep 16 12:24:45 2013 From: chris at ambigc.com (Chris Hamilton) Date: Mon, 16 Sep 2013 04:24:45 -0700 Subject: [winswitch] Terminal emulators with 0.12.19 Message-ID: <0d0e01ceb2cf$5a2c2520$0e846f60$@ambigc.com> Hi, I have tried a few 'fancy' terminal emulators - terminator, konsole, and gterm and they all seem unable to refresh or display a consistent console window. Either the bottom refuses to display any text updates, sections in the top or middle sometimes do not display, though drag highlighting text with the mouse refreshes. Resizing the window causes scroll bar and other element artifacts to not redraw. xterm works perfectly fine, other than it is xterm. I haven't found anything in the documentation or forums to help, so I thought I would ask if anyone else has seen this and any suggestions. This is an Ubuntu raring winswitch 0.12.19 server from the winswitch Ubuntu repo and a windows 7 0.12.19 client. Thanks, Chris Hamilton From antoine at nagafix.co.uk Mon Sep 16 13:22:48 2013 From: antoine at nagafix.co.uk (Antoine Martin) Date: Mon, 16 Sep 2013 19:22:48 +0700 Subject: [winswitch] Terminal emulators with 0.12.19 In-Reply-To: <0d0e01ceb2cf$5a2c2520$0e846f60$@ambigc.com> References: <0d0e01ceb2cf$5a2c2520$0e846f60$@ambigc.com> Message-ID: <5236F818.7070608@nagafix.co.uk> On 16/09/13 18:24, Chris Hamilton wrote: > Hi, I have tried a few 'fancy' terminal emulators - terminator, konsole, and > gterm and they all seem unable to refresh or display a consistent console > window. Either the bottom refuses to display any text updates, sections in > the top or middle sometimes do not display, though drag highlighting text > with the mouse refreshes. Resizing the window causes scroll bar and other > element artifacts to not redraw. xterm works perfectly fine, other than it > is xterm. I haven't found anything in the documentation or forums to help, > so I thought I would ask if anyone else has seen this and any suggestions. > This is an Ubuntu raring winswitch 0.12.19 server from the winswitch Ubuntu > repo and a windows 7 0.12.19 client. This could be related to this bug (sounds similar): https://www.xpra.org/trac/ticket/429 Even if not, some of the suggestions in the ticket are worth a try. (disable GL, ..) Cheers Antoine From varvello at yahoo.com Thu Sep 19 13:20:30 2013 From: varvello at yahoo.com (Davide Varvello) Date: Thu, 19 Sep 2013 05:20:30 -0700 (PDT) Subject: [winswitch] Keyexpired during update Message-ID: <1379593230.61596.YahooMailNeo@web160104.mail.bf1.yahoo.com> Hi Antoine, ?During an apt-get update I receive a warning: "W: GPG error: http://winswitch.org lucid Release: The following signatures were invalid: KEYEXPIRED 1367717527" Is it something I have to worry about? Cheers ?Davide From antoine at nagafix.co.uk Thu Sep 19 13:27:18 2013 From: antoine at nagafix.co.uk (Antoine Martin) Date: Thu, 19 Sep 2013 19:27:18 +0700 Subject: [winswitch] Keyexpired during update In-Reply-To: <1379593230.61596.YahooMailNeo@web160104.mail.bf1.yahoo.com> References: <1379593230.61596.YahooMailNeo@web160104.mail.bf1.yahoo.com> Message-ID: <523AEDA6.9040808@nagafix.co.uk> On 19/09/13 19:20, Davide Varvello wrote: > Hi Antoine, > During an apt-get update I receive a warning: "W: GPG error: http://winswitch.org lucid Release: The following signatures were invalid: KEYEXPIRED 1367717527" > > Is it something I have to worry about? Support for Lucid was dropped a while back. You should upgrade. As for the key, there is a new one and this is covered in the FAQ: http://xpra.org/trac/wiki/FAQ#WarningsandMessages Antoine From thomas at koch.ro Fri Sep 20 07:39:20 2013 From: thomas at koch.ro (Thomas Koch) Date: Fri, 20 Sep 2013 08:39:20 +0200 Subject: [winswitch] xpra list is empty immediately after xpra start Message-ID: <201309200839.20957.thomas@koch.ro> Hi, I've made a script that is called on every ssh login to my virtual machine. It tests whether xpra is running (via xpra list) and if not it starts xpra and immediately afterwards it runs xpra list to parse the value that should be assigned to $DISPLAY. It's always(?) the case, that xpra list does not show the just started instance but prints an empty list. Do you consider this a bug? Maybe you could add an option to xpra start that it should only return once everything is set up so that xpra list finds the instance? While I'm at it: It would be nice to have a shell parsable output of xpra list, e.g. separated by tabs without indentation. Regards, Thomas Koch From antoine at nagafix.co.uk Fri Sep 20 07:46:02 2013 From: antoine at nagafix.co.uk (Antoine Martin) Date: Fri, 20 Sep 2013 13:46:02 +0700 Subject: [winswitch] xpra list is empty immediately after xpra start In-Reply-To: <201309200839.20957.thomas@koch.ro> References: <201309200839.20957.thomas@koch.ro> Message-ID: <523BEF2A.20709@nagafix.co.uk> On 20/09/13 13:39, Thomas Koch wrote: > Hi, > > I've made a script that is called on every ssh login to my virtual machine. It > tests whether xpra is running (via xpra list) and if not it starts xpra and > immediately afterwards it runs xpra list to parse the value that should be > assigned to $DISPLAY. > > It's always(?) the case, that xpra list does not show the just started > instance but prints an empty list. > > Do you consider this a bug? Maybe you could add an option to xpra start that > it should only return once everything is set up so that xpra list finds the > instance? This is not considered a bug at present and is probably not (easily) fixable either. That's because servers daemonize before setting up everything. > While I'm at it: It would be nice to have a shell parsable output of xpra > list, e.g. separated by tabs without indentation. Parsing spaces isn't difficult, and changing the output may break existing scripts.. Antoine From varvello at yahoo.com Sat Sep 21 06:59:48 2013 From: varvello at yahoo.com (Davide Varvello) Date: Fri, 20 Sep 2013 22:59:48 -0700 (PDT) Subject: [winswitch] Keyexpired during update In-Reply-To: <523AEDA6.9040808@nagafix.co.uk> References: <1379593230.61596.YahooMailNeo@web160104.mail.bf1.yahoo.com> <523AEDA6.9040808@nagafix.co.uk> Message-ID: <1379743188.64769.YahooMailNeo@web160105.mail.bf1.yahoo.com> Hi Martin, ?I imported the key (according to instructions here: http://www.madboa.com/geek/gpg-quickstart/), but when I do an apt-get I receive the same warning. Davide >________________________________ > From: Antoine Martin >To: shifter-users at lists.devloop.org.uk >Sent: Thursday, September 19, 2013 2:27 PM >Subject: Re: [winswitch] Keyexpired during update > > >On 19/09/13 19:20, Davide Varvello wrote: >> Hi Antoine, >>? During an apt-get update I receive a warning: "W: GPG error: http://winswitch.org lucid Release: The following signatures were invalid: KEYEXPIRED 1367717527" >> >> Is it something I have to worry about? >Support for Lucid was dropped a while back. You should upgrade. >As for the key, there is a new one and this is covered in the FAQ: >http://xpra.org/trac/wiki/FAQ#WarningsandMessages > >Antoine >_______________________________________________ >shifter-users mailing list >shifter-users at lists.devloop.org.uk >http://lists.devloop.org.uk/mailman/listinfo/shifter-users > > > From chris at ambigc.com Sun Sep 22 05:31:43 2013 From: chris at ambigc.com (Chris Hamilton) Date: Sat, 21 Sep 2013 21:31:43 -0700 Subject: [winswitch] Terminal emulators with 0.12.19 update Message-ID: <002401ceb74c$a5c4d1e0$f14e75a0$@ambigc.com> Hi, sorry I am not in the mail list, so this will not be in the right thread. I followed directions in https://www.xpra.org/trac/ticket/429 however they did not seem to work. I had similar issues manually setting XPRA_ALLOW_ALPHA as well as various command line parameters on xpra in Ubuntu, but nothing changed. OpenGL on client or server didn't seem to change anything either. I eventually set the xpra.conf encoding default to png in Windows and the terminals finally worked completely. From antoine at nagafix.co.uk Sun Sep 22 15:19:49 2013 From: antoine at nagafix.co.uk (Antoine Martin) Date: Sun, 22 Sep 2013 21:19:49 +0700 Subject: [winswitch] Keyexpired during update In-Reply-To: <1379743188.64769.YahooMailNeo@web160105.mail.bf1.yahoo.com> References: <1379593230.61596.YahooMailNeo@web160104.mail.bf1.yahoo.com> <523AEDA6.9040808@nagafix.co.uk> <1379743188.64769.YahooMailNeo@web160105.mail.bf1.yahoo.com> Message-ID: <523EFC85.3050500@nagafix.co.uk> On 21/09/13 12:59, Davide Varvello wrote: > Hi Martin, > I imported the key (according to instructions here: > http://www.madboa.com/geek/gpg-quickstart/), but when I do an apt-get > I receive the same warning. I don't know what you did exactly, so I can't help you. As per the FAQ, delete the key and re-import it (as when you first installed the repository) Antoine > > Davide > > > ------------------------------------------------------------------------ > *From:* Antoine Martin > *To:* shifter-users at lists.devloop.org.uk > *Sent:* Thursday, September 19, 2013 2:27 PM > *Subject:* Re: [winswitch] Keyexpired during update > > On 19/09/13 19:20, Davide Varvello wrote: > > Hi Antoine, > > During an apt-get update I receive a warning: "W: GPG error: > http://winswitch.org lucid Release: The > following signatures were invalid: KEYEXPIRED 1367717527" > > > > Is it something I have to worry about? > Support for Lucid was dropped a while back. You should upgrade. > As for the key, there is a new one and this is covered in the FAQ: > http://xpra.org/trac/wiki/FAQ#WarningsandMessages > > Antoine > _______________________________________________ > shifter-users mailing list > shifter-users at lists.devloop.org.uk > > http://lists.devloop.org.uk/mailman/listinfo/shifter-users > > From antoine at nagafix.co.uk Sun Sep 22 15:25:17 2013 From: antoine at nagafix.co.uk (Antoine Martin) Date: Sun, 22 Sep 2013 21:25:17 +0700 Subject: [winswitch] Terminal emulators with 0.12.19 update In-Reply-To: <002401ceb74c$a5c4d1e0$f14e75a0$@ambigc.com> References: <002401ceb74c$a5c4d1e0$f14e75a0$@ambigc.com> Message-ID: <523EFDCD.6090200@nagafix.co.uk> On 22/09/13 11:31, Chris Hamilton wrote: > Hi, sorry I am not in the mail list, so this will not be in the right > thread. > > > > I followed directions in https://www.xpra.org/trac/ticket/429 however they > did not seem to work. I had similar issues manually setting > XPRA_ALLOW_ALPHA as well as various command line parameters on xpra in > Ubuntu, but nothing changed. OpenGL on client or server didn't seem to > change anything either. I eventually set the xpra.conf encoding default to > png in Windows and the terminals finally worked completely. I don't really know what problem you are referring to... so the answer below is just a guess: If you use an encoding that does not use swscale (all except x264 and vpx), then you may bypass problems with broken builds. However, all the other encodings are also much much less efficient (especially PNG), and xpra is now designed to run with the newer/better encodings. So, I still recommend that you resolve your swscale/x264 library issues. Antoine From rbees.construction at gmail.com Sun Sep 22 01:46:48 2013 From: rbees.construction at gmail.com (Randy) Date: Sat, 21 Sep 2013 20:46:48 -0400 Subject: [winswitch] Keyexpired during update In-Reply-To: <1379743188.64769.YahooMailNeo@web160105.mail.bf1.yahoo.com> References: <1379593230.61596.YahooMailNeo@web160104.mail.bf1.yahoo.com> <523AEDA6.9040808@nagafix.co.uk> <1379743188.64769.YahooMailNeo@web160105.mail.bf1.yahoo.com> Message-ID: <10539838.m44TvSk49H@bumblebee> I reciently had this problem too. It was caused by upgrading to the new testing (Jessie) after Wheezy went stable and severl keys were changed. What I had to do: 1. comment out the repo entry in souces.list 2. run apt-get update 3. locate the offending key for your distro and version and delete it. I forget the exact commands or where they are actually stored as it has been a few months now. 4. re install the new key and run apt-get update 5. re enable the repo entry in sources.list and run apt-get update again Hope this helps On Friday, September 20, 2013 10:59:48 PM Davide Varvello wrote: > Hi Martin, > I imported the key (according to instructions here: > http://www.madboa.com/geek/gpg-quickstart/), but when I do an apt-get I > receive the same warning. > > Davide > > >________________________________ > > > > From: Antoine Martin > > > >To: shifter-users at lists.devloop.org.uk > >Sent: Thursday, September 19, 2013 2:27 PM > >Subject: Re: [winswitch] Keyexpired during update > > > >On 19/09/13 19:20, Davide Varvello wrote: > >> Hi Antoine, > >> > >> During an apt-get update I receive a warning: "W: GPG error: > >>http://winswitch.org lucid Release: The following signatures were > >>invalid: KEYEXPIRED 1367717527">> > >> Is it something I have to worry about? > > > >Support for Lucid was dropped a while back. You should upgrade. > >As for the key, there is a new one and this is covered in the FAQ: > >http://xpra.org/trac/wiki/FAQ#WarningsandMessages > > > >Antoine > >_______________________________________________ > >shifter-users mailing list > >shifter-users at lists.devloop.org.uk > >http://lists.devloop.org.uk/mailman/listinfo/shifter-users > > _______________________________________________ > shifter-users mailing list > shifter-users at lists.devloop.org.uk > http://lists.devloop.org.uk/mailman/listinfo/shifter-users -- Randy Batterbee Take a trip with Torah and find out who you really are. Open Source & Linux if it ain't broke tweak it you can always reinstall and it's free From antoine at nagafix.co.uk Sun Sep 22 15:31:18 2013 From: antoine at nagafix.co.uk (Antoine Martin) Date: Sun, 22 Sep 2013 21:31:18 +0700 Subject: [winswitch] Keyexpired during update In-Reply-To: <10539838.m44TvSk49H@bumblebee> References: <1379593230.61596.YahooMailNeo@web160104.mail.bf1.yahoo.com> <523AEDA6.9040808@nagafix.co.uk> <1379743188.64769.YahooMailNeo@web160105.mail.bf1.yahoo.com> <10539838.m44TvSk49H@bumblebee> Message-ID: <523EFF36.5060906@nagafix.co.uk> On 22/09/13 07:46, Randy wrote: > I reciently had this problem too. It was caused by upgrading to the new > testing (Jessie) after Wheezy went stable and severl keys were changed. > > What I had to do: > 1. comment out the repo entry in souces.list AFAIK, This is not needed. > 2. run apt-get update Not needed. > 3. locate the offending key for your distro and version and delete it. I > forget the exact commands or where they are actually stored as it has > been a few months now. It's the key from the error message: 1367717527 > 4. re install the new key and run apt-get update Yes. > 5. re enable the repo entry in sources.list and run apt-get update again AFAIK, this is not needed. If you want to force the refresh, you can "apt-get clean && apt-get update" From thejunk.b at gmail.com Sun Sep 22 17:12:14 2013 From: thejunk.b at gmail.com (Randy) Date: Sun, 22 Sep 2013 12:12:14 -0400 Subject: [winswitch] Keyexpired during update In-Reply-To: <523EFF36.5060906@nagafix.co.uk> References: <1379593230.61596.YahooMailNeo@web160104.mail.bf1.yahoo.com> <10539838.m44TvSk49H@bumblebee> <523EFF36.5060906@nagafix.co.uk> Message-ID: <3609709.ayQezIHZqM@bumblebee> On Sunday, September 22, 2013 09:31:18 PM Antoine Martin wrote: > On 22/09/13 07:46, Randy wrote: > > I reciently had this problem too. It was caused by upgrading to the new > > testing (Jessie) after Wheezy went stable and severl keys were changed. > > > > What I had to do: > > 1. comment out the repo entry in souces.list > > AFAIK, This is not needed. > > > 2. run apt-get update > > Not needed. > > > 3. locate the offending key for your distro and version and delete it. I > > forget the exact commands or where they are actually stored as it has > > been a few months now. > > It's the key from the error message: 1367717527 > > > 4. re install the new key and run apt-get update > > Yes. > > > 5. re enable the repo entry in sources.list and run apt-get update again > > AFAIK, this is not needed. > If you want to force the refresh, you can "apt-get clean && apt-get update" Actually all these steps were needed to resoleve the error for the Virtualbox repo after the update to the new Debian Testing (Jessie). True with winswitch/xpra the issue was resolved by just importing the new key and refreshing the repo. So the problem I had was probably related to my system and not an issue in general. Thanks _______________________________________________ > shifter-users mailing list > shifter-users at lists.devloop.org.uk > http://lists.devloop.org.uk/mailman/listinfo/shifter-users -- If it ain't broke tweek it From varvello at yahoo.com Wed Sep 25 08:27:25 2013 From: varvello at yahoo.com (Davide Varvello) Date: Wed, 25 Sep 2013 00:27:25 -0700 (PDT) Subject: [winswitch] Keyexpired during update In-Reply-To: <523EFC85.3050500@nagafix.co.uk> References: <1379593230.61596.YahooMailNeo@web160104.mail.bf1.yahoo.com> <523AEDA6.9040808@nagafix.co.uk> <1379743188.64769.YahooMailNeo@web160105.mail.bf1.yahoo.com> <523EFC85.3050500@nagafix.co.uk> Message-ID: <1380094045.19181.YahooMailNeo@web160101.mail.bf1.yahoo.com> I solved. I thought the command:?gpg --import?yourKey.asc?(as shown?here:?http://www.madboa.com/geek/gpg-quickstart/) was the right way, instead I had to write: apt-key add yourKey.asc? as pointed here:?http://askubuntu.com/questions/291035/how-to-add-a-gpg-key-to-the-apt-sources-keyring Davide >________________________________ > > > >On 21/09/13 12:59, Davide Varvello wrote: > >Hi Martin, >>?I imported the key (according to instructions here: http://www.madboa.com/geek/gpg-quickstart/), but when I do an apt-get I receive the same warning. I don't know what you did exactly, so I can't help you. >As per the FAQ, delete the key and re-import it (as when you first installed the repository) > >Antoine > > > >> >>Davide >> >> >> >> >> >>>________________________________ >>> >>>On 19/09/13 19:20, Davide Varvello wrote: >>>> Hi Antoine, >>>>? During an apt-get update I receive a warning: "W: GPG error: http://winswitch.org lucid Release: The following signatures were invalid: KEYEXPIRED 1367717527" >>>> >>>> Is it something I have to worry about? >>>Support for Lucid was dropped a while back. You should upgrade. >>>As for the key, there is a new one and this is covered in the FAQ: >>>http://xpra.org/trac/wiki/FAQ#WarningsandMessages >>> >>>Antoine >>>_______________________________________________ >>>shifter-users mailing list >>>shifter-users at lists.devloop.org.uk >>>http://lists.devloop.org.uk/mailman/listinfo/shifter-users >>> >>> >>> > > > From nick at sqrt.co.uk Wed Sep 25 11:24:48 2013 From: nick at sqrt.co.uk (Nick Burrett) Date: Wed, 25 Sep 2013 11:24:48 +0100 Subject: [winswitch] Terminal emulators with 0.12.19 update In-Reply-To: <523EFDCD.6090200@nagafix.co.uk> References: <002401ceb74c$a5c4d1e0$f14e75a0$@ambigc.com> <523EFDCD.6090200@nagafix.co.uk> Message-ID: In my experience, the encodings that I find to work the best are: PNG, PNG/P and PNG/L. PNG I use on a local GigE LAN. PNG/L and PNG/P with compression level 9. I use for emacs/PyCharm/gnome-terminal where my client is in Ireland and the server is in North Sweden (50-60ms away). x264 appears to have a higher encoding latency and tends to make screen updates jumpy. My server has 24 CPUs and the client 4 CPUs, so there should be enough CPU capacity to encode at a decent rate, if threaded sufficiently. On 22 September 2013 15:25, Antoine Martin wrote: > On 22/09/13 11:31, Chris Hamilton wrote: >> Hi, sorry I am not in the mail list, so this will not be in the right >> thread. >> >> >> >> I followed directions in https://www.xpra.org/trac/ticket/429 however they >> did not seem to work. I had similar issues manually setting >> XPRA_ALLOW_ALPHA as well as various command line parameters on xpra in >> Ubuntu, but nothing changed. OpenGL on client or server didn't seem to >> change anything either. I eventually set the xpra.conf encoding default to >> png in Windows and the terminals finally worked completely. > I don't really know what problem you are referring to... so the answer > below is just a guess: > If you use an encoding that does not use swscale (all except x264 and > vpx), then you may bypass problems with broken builds. > However, all the other encodings are also much much less efficient > (especially PNG), and xpra is now designed to run with the newer/better > encodings. > So, I still recommend that you resolve your swscale/x264 library issues. > > Antoine > _______________________________________________ > shifter-users mailing list > shifter-users at lists.devloop.org.uk > http://lists.devloop.org.uk/mailman/listinfo/shifter-users From antoine at nagafix.co.uk Wed Sep 25 12:58:27 2013 From: antoine at nagafix.co.uk (Antoine Martin) Date: Wed, 25 Sep 2013 18:58:27 +0700 Subject: [winswitch] encodings and compression [was: Terminal emulators with 0.12.19 update] In-Reply-To: References: <002401ceb74c$a5c4d1e0$f14e75a0$@ambigc.com> <523EFDCD.6090200@nagafix.co.uk> Message-ID: <5242CFE3.20102@nagafix.co.uk> On 25/09/13 17:24, Nick Burrett wrote: > In my experience, the encodings that I find to work the best are: PNG, > PNG/P and PNG/L. PNG I use on a local GigE LAN. On a Gigabit LAN or better, isn't plain RGB+zlib better? Even playing an HD video should not saturate the link, and PNG is unlikely to be able to cope with 30fps. > PNG/L and PNG/P with > compression level 9. If you can tolerate the limited number of colours of PNG/L or PNG/P, they will beat everything else which uses true colour encoding and therefore has to work a lot harder. Compression level 9 does absolutely nothing for picture compression size, and only adds to the latency for the rest of the packets for very little space saving. You should really avoid it. FYI: trunk adds lz4 support for -z=1, it is faster than zlib (but compresses a little less - which doesn't matter much) > I use for emacs/PyCharm/gnome-terminal where my > client is in Ireland and the server is in North Sweden (50-60ms away). > x264 appears to have a higher encoding latency and tends to make > screen updates jumpy. If you care about latency, you should set the speed option high enough. FYI: trunk is likely to get hardware assisted h264 encoding support soon (using NVidia's NVENC), which should give us H264 HQ encoding with very low latency on supported hardware. > My server has 24 CPUs and the client 4 CPUs, so > there should be enough CPU capacity to encode at a decent rate, if > threaded sufficiently. The client CPU capacity rarely matters - unless you are on an embedded CPU (ARM etc). x264 is partially threaded (in recent versions at least), none of the other are - and PNG is a CPU pig. On top of that, PNG cannot be tuned, whereas x264 can (as per comment about speed above). It is possible for PNG to be faster (lower latency, but often lower compression too) than x264 when used at a low enough framerate - but for anything else, it is generally a poor option. (bar special case of PNG/L and PNG/P) To ascertain which encodings are the best for which application at which resolution on which hardware, there are scripts in the source tree which can run for hours and produce CSV output for analysis. The comments above are based on weeks of testing (man-hours), but using a limited set of applications and settings - it is possible, likely even, that other workloads will behave differently. If you have specific applications which behave poorly, we can look at them and try to improve things. Antoine > > > On 22 September 2013 15:25, Antoine Martin wrote: >> On 22/09/13 11:31, Chris Hamilton wrote: >>> Hi, sorry I am not in the mail list, so this will not be in the right >>> thread. >>> >>> >>> >>> I followed directions in https://www.xpra.org/trac/ticket/429 however they >>> did not seem to work. I had similar issues manually setting >>> XPRA_ALLOW_ALPHA as well as various command line parameters on xpra in >>> Ubuntu, but nothing changed. OpenGL on client or server didn't seem to >>> change anything either. I eventually set the xpra.conf encoding default to >>> png in Windows and the terminals finally worked completely. >> I don't really know what problem you are referring to... so the answer >> below is just a guess: >> If you use an encoding that does not use swscale (all except x264 and >> vpx), then you may bypass problems with broken builds. >> However, all the other encodings are also much much less efficient >> (especially PNG), and xpra is now designed to run with the newer/better >> encodings. >> So, I still recommend that you resolve your swscale/x264 library issues. >> >> Antoine >> _______________________________________________ >> shifter-users mailing list >> shifter-users at lists.devloop.org.uk >> http://lists.devloop.org.uk/mailman/listinfo/shifter-users