dvdhrm / kmscon Goto Github PK

when I run kmscon -l /bin/login and press backspace while typing in login prompt ^H appears. eta^Ham is not the same as etam. This also applies to typing password (like whole password, nothing appears this time, but you cannot correct yourself). In bash backspace behaves as expected.

kmscon(1) references kmscon.conf(5), which doesn't exist.

kmscon ignores system global fontconfig settings. (/etc/fonts/...) while fbterm(another terminal emulator than kmscon) respects it.

I wanted kmscon to use NanumGothicCoding font but when it is launched by /etc/inittab it uses another ***ing ugly font, though it uses NanumGothicCoding when I launch it in agetty by typing "kmscon" while there is ~/.fonts.conf.

my /etc/inittab is written like :

# TERMINALS
c1:12345:respawn:/sbin/agetty 38400 tty1 linux
#c2:2345:respawn:/sbin/agetty 38400 tty2 linux
#c3:2345:respawn:/sbin/agetty 38400 tty3 linux
#c4:2345:respawn:/sbin/agetty 38400 tty4 linux
#c5:2345:respawn:/sbin/agetty 38400 tty5 linux
#c6:2345:respawn:/sbin/agetty 38400 tty6 linux
#c1:12345:respawn:/usr/bin/kmscon --vt /dev/tty1
c2:2345:respawn:/usr/bin/kmscon --vt /dev/tty2
c3:2345:respawn:/usr/bin/kmscon --vt /dev/tty3
c4:2345:respawn:/usr/bin/kmscon --vt /dev/tty4
c5:2345:respawn:/usr/bin/kmscon --vt /dev/tty5
c6:2345:respawn:/usr/bin/kmscon --vt /dev/tty6

my /etc/fonts/conf.d/98-ko-dewr.conf is :

<?xml version="1.0"?>
<!DOCTYPE fontconfig SYSTEM "fonts.dtd">
<fontconfig>
        <!--
                Private font directory
        <dir prefix="xdg">fonts</dir> -->

        <!-- Generic names -->
        <alias>
                <family>sans-serif</family>
                <prefer>
                        <family>UnDotum</family>
                </prefer>
        </alias>
        <alias>
                <family>serif</family>
                <prefer>
                        <family>UnBatang</family>
                </prefer>
        </alias>
        <alias>
                <family>monospace</family>
                <prefer>
                        <family>NanumGothicCoding</family>
                </prefer>
        </alias>

        <!-- disable hinting -->
<!--
        <match target="font">
                <test name="family">
                        <string>IPAPGothic</string>
                        <string>IPAPMincho</string>
                        <string>IPAGothic</string>
                        <string>IPAMincho</string>
                </test>
                <edit name="hinting" mode="assign">
                        <bool>false</bool>
                </edit>
        </match>
-->
<!--
        use UnBatang font when serif is requested for Korean
-->
<match>
        <test name="lang" compare="contains">
                <string>ko</string>
        </test>
        <test name="family">
                <string>serif</string>
        </test>
        <edit name="family" mode="prepend">
                <string>UnBatang</string>
        </edit>
</match>

<!--
        use UnDotum font when sans-serif is requested for Korean
-->
<match>
        <test name="lang" compare="contains">
                <string>ko</string>
        </test>
        <test name="family">
                <string>sans-serif</string>
        </test>
        <edit name="family" mode="prepend">
                <string>UnDotum</string>
        </edit>
</match>

<!--
        use NanumGothicCoding font when mono is requested for Korean
-->
<match>
        <test name="lang" compare="contains">
                <string>ko</string>
        </test>
        <test name="family">
                <string>monospace</string>
        </test>
        <edit name="family" mode="prepend">
                <string>NanumGothicCoding</string>
        </edit>
</match>

</fontconfig>

test environment : a gentoo box

My build options are:

CFLAGS="-Os -static" LDFLAGS="-static" ./configure --disable-wlterm --disable-silent-rules --disable-multi-seat --with-video=fbdev --with-renderers=bbulk --with-fonts=unifont --with-sessions=dummy,terminal

But the binary after finishing build still have dynamic linked libs:

Tag Type Name/Value
0x00000001 (NEEDED) Shared library: [libudev.so.1]
0x00000001 (NEEDED) Shared library: [libpciaccess.so.0]
0x00000001 (NEEDED) Shared library: [libpthread.so.0]
0x00000001 (NEEDED) Shared library: [libxkbcommon.so.0]
0x00000001 (NEEDED) Shared library: [libgcc_s.so.1]
0x00000001 (NEEDED) Shared library: [libc.so.6]

Only the binary is much bigger ~15MB. Only the libgcc_s.so doesn't exist as static lib on my system, all other are have .a files (udev too, I have build systemd with enabled static libs)

Chinese characters are usually as twice wide as the Latin characters. It seems currently kmscon assumes that every character can be rendered in a space as wide as any ASCII character. This leads to a bug that only half of any wide-character is shown.

here is the related code in text_font_pango.c:
in manager_get_face:
str = "abcdefghijklmnopqrstuvwxyz"
"ABCDEFGHIJKLMNOPQRSTUVWXYZ"
"@!"$%&/()=?}][{°^~+*#'<>|-_.:,;`´";
num = strlen(str);
pango_layout_set_text(layout, str, num);
pango_layout_get_pixel_extents(layout, NULL, &rec);

memcpy(&face->real_attr, &face->attr, sizeof(face->attr));
face->real_attr.height = rec.height;
face->real_attr.width = rec.width / num + 1;

and in get_glyph

pango_layout_line_get_pixel_extents(line, NULL, &rec);
glyph->buf.width = face->real_attr.width;
glyph->buf.height = face->real_attr.height;

xterm (and compatible) have a feature called "Alternate Screen Buffer" that's used by most "full screen" terminal applications, like vi, less, man, mutt and screen - it's in the default termcap init/deinit called ti/te or rmcup/smcup. When the application starts, it switches to an alternate screen buffer, which doesn't have a scroll buffer, and the original screen is left untouched. When the application quits/suspends, the original buffer is restored as if the the application never existed on the terminal. There's only one alternate screen, they're not stacked or something like that.

Some application allow to disable this, because some people don't like it (vim has t_ti and t_te, less has LESS=X, and xterm has the titeInhibit resource). Also some terminal emulators like the linux VT don't support it at all.

Someone did an illustration here:
http://fixlog.blogspot.co.il/2006/09/stop-gnome-terminal-screen-clear.html

Here's the xterm reference:
http://invisible-island.net/xterm/ctlseqs/ctlseqs.html#The Alternate Screen Buffer
The relevant DEC control sequences are 47, 1047, 1048, 1049. Applications only use 1049 as far as I can tell.

Here's the xterm source code:
http://anonscm.debian.org/gitweb/?p=pkg-xorg/app/xterm.git;a=blob;f=charproc.c;h=ec5f498d714615595832a3764d31f8f77232fe54;hb=HEAD#l4340

This feature is a bit annoying to implement, I think, but should basically boil down to toggling two tsm_screen's. When I tried it I got into some trouble, but with some refactoring it seems doable.

Any plans for a way to specify a tty at startup?

This is necessary to use a [email protected] for systemd's autovt service.

following up issue #17:

Regarding speed: What graphics card are you using? gltex should be way faster than any other renderer if you have a modern graphics card.

Intel GM45 on a ThinkPad X301. I think this called gen4.

On low-end machines (like my Intel-Atom N450 with integrated i915) the bblit renderer is still faster, indeed. The reasons why kernel-console or other consoles are faster is probably that they use partial-update-renderers (what I call them). That is, they redraw only the part of the screen that changed. kmscon however, always redraws the whole screen.

This cannot be the whole story. So I compare kmscon and fbterm. fbterm is faster in all respects. time dmesg is 0.26s on kmscon and 0.13s on fbterm, twice as fast. It’s quite striking for full-screen thingies that draw every character. There’s a package called caca-utils, that has cacademo and cacafire, which should be instructive. Mplayer also has a caca output module.
This is superslow in kmscon, it pegs one core only for kmscon.

Since fbterm works like the dumb fallback of kmscon it should not be so much faster.

Also the fbcon standard console is faster in one respect: for fullscreen drawing with character addressing. For normal scrolling like dmesg it is indeed slower: time dmesg is 1.9s

The reason is, that this ensures that I can realiably measure delays. In other words, the console will never be slower under any circumstances. If I implement partial-redraws (which I do intend!) I can never be sure that under other use-cases the terminal may be slower.

But the massive slowdown happens when every character is updated anyway, not for linear scrolling.

Build fails with "src/conf.c:921:2: error: too few arguments to function 'xkb_keysym_from_name'"
http://lists.freedesktop.org/archives/wayland-devel/2012-October/005885.html

In README there is "libxkbcommon has no public release" but it has changed: libxkbcommon-0.2.0 was released http://cgit.freedesktop.org/xorg/lib/libxkbcommon/tag/?id=xkbcommon-0.2.0

I've been putting off asking about this because the answer is probably obvious or something stupid that I'm missing.

When I'm in kmscon, I don't get bold characters. For example, notice and error messages from journalctl are not bold. Also, with lsblk and findmnt I do get the unicode line drawing characters, but they're the thinner alternatives.

kmscon won't draw underlines. I tested it in a gentoo box with manpage(less), vim and w3m

"it produces "^H" when I hit [backspace] on kmscon login phase" instead of removing a character
in my gentoo box

With default configure options, there is not kmscon binary created that I can see.

The only binaries I see created are genshader* & genunifont*.

Finding and trying to use the ./configure --enable-kmscon option results in:

checking which render backends the user wants... bblit,bbulk,gltex (default)
checking which font backends the user wants... 8x16,freetype2,pango (default)
checking which sessions the user wants... dummy,terminal,cdev (default)
./configure: line 14329: xno: command not found
configure: error: eloop, TSM or uterm cannot be built for kmscon

Suppose my --xkb-layout is "us,il", and the active layout is "il". This layout produces non-ASCII keysyms in the base level. In this case typing e.g. Ctrl+l (that is the lower case 'l' keyboard key / xkb keycode) should still clear the screen, etc. This works in xterm, urxvt and libvte (though strangely Ctrl is the only modifier which gets this special treatment).

libvte had these bugs about this issue:
https://bugzilla.gnome.org/show_bug.cgi?id=375112
https://bugzilla.gnome.org/show_bug.cgi?id=589557

They solved it here:
http://git.gnome.org/browse/vte/commit/?id=d64e28b3290aaf034ec37bbf4524ef78adbdfb2e
Which claims xterm does the same. It seem to work, so I think kmscon should do the same.

The equivalent loop in xkbcommon would look something like this:
https://gist.github.com/3833422
This is simple enough, and I've made sure it works as expected (it requires some relatively new API though). However the tsm_vte does not facilitate doing this directly like the libvte code above does (it doesn't have access to the xkb_state), so I didn't attempt a patch myself.

Because kmscon dynamically allocates pty devices it makes /etc/securetty useless as an access control mechanism. Being able to make access decisions based on "someone is on a local console" is a useful feature, so it would be nice if there were some way to expose this information to PAM in the kmscon environment.

I currently use fbsplash to set a background image for virtual terminals.
But fbsplash is not maintained anymore and can't be compiled in new kernels.

I realized kmscon would be a good replacement if it supported background images.

I have an issue where I start kmscon, where the session is something like this:

$ ls /
/ /home for bar baz

And then the prompt doesn't appear again. VT switching still works, so it's not some infinite loop. I bisected this to commit a0c644f which changed the default TERM to xterm-256color (but it also happens with just xterm). It doesn't happen with TERM=vt220.

After switching to another VT and attaching gdb to kmscon, the bt is:
#0 0xb773b424 in __kernel_vsyscall ()
#1 0xb747ee78 in __epoll_wait_nocancel () from /usr/lib/libc.so.6
#2 0xb7716a6e in ev_eloop_dispatch (loop=loop@entry=0x9e460c0, timeout=timeout@entry=-1) at src/eloop.c:808
#3 0xb771705e in ev_eloop_run (loop=0x9e460c0, timeout=timeout@entry=-1) at src/eloop.c:895
#4 0x0804b53d in main (argc=1, argv=0xbfe78e84) at src/main.c:624

So nothing helpful here. When attaching to the bash process, the bt is:
#0 0xb77b6424 in __kernel_vsyscall ()
#1 0xb763b6b3 in __read_nocancel () from /usr/lib/libc.so.6
#2 0xb7788b84 in rl_getc () from /usr/lib/libreadline.so.6
#3 0xb778941b in rl_read_key () from /usr/lib/libreadline.so.6
#4 0xb777302c in readline_internal_char () from /usr/lib/libreadline.so.6
#5 0xb7773585 in readline () from /usr/lib/libreadline.so.6
#6 0x0805cdca in ?? ()
#7 0x0805e90e in ?? ()
#8 0x080614e3 in ?? ()
#9 0x08064622 in yyparse ()
#10 0x0805c6cd in parse_command ()
#11 0x0805c78d in read_command ()
#12 0x0805c9d7 in reader_loop ()
#13 0x0805acfa in main ()

So it just waits for user input as well. So I don't know where to look (I can bisect further back with setting --term manually if you don't have a quick idea).

Since it clearly doesn't happen to other people, some details of this machine:

uname -a:
Linux fst 3.5.3-1-ARCH #1 SMP PREEMPT Sun Aug 26 08:15:06 UTC 2012 i686 GNU/Linux

packages:
extra/mesa 8.0.4-3 [installed]
extra/libdrm 2.4.39-1 [installed]
extra/intel-dri 8.0.4-3 [installed]

glxinfo:
name of display: :0.0
display: :0 screen: 0
direct rendering: Yes
server glx vendor string: SGI
server glx version string: 1.4
server glx extensions:
GLX_ARB_multisample, GLX_EXT_import_context, GLX_EXT_texture_from_pixmap,
GLX_EXT_visual_info, GLX_EXT_visual_rating, GLX_MESA_copy_sub_buffer,
GLX_OML_swap_method, GLX_SGI_swap_control, GLX_SGIS_multisample,
GLX_SGIX_fbconfig, GLX_SGIX_pbuffer, GLX_SGIX_visual_select_group,
GLX_INTEL_swap_event
client glx vendor string: Mesa Project and SGI
client glx version string: 1.4
client glx extensions:
GLX_ARB_create_context, GLX_ARB_create_context_profile,
GLX_ARB_get_proc_address, GLX_ARB_multisample, GLX_EXT_import_context,
GLX_EXT_visual_info, GLX_EXT_visual_rating, GLX_EXT_framebuffer_sRGB,
GLX_EXT_create_context_es2_profile, GLX_MESA_copy_sub_buffer,
GLX_MESA_multithread_makecurrent, GLX_MESA_swap_control,
GLX_OML_swap_method, GLX_OML_sync_control, GLX_SGI_make_current_read,
GLX_SGI_swap_control, GLX_SGI_video_sync, GLX_SGIS_multisample,
GLX_SGIX_fbconfig, GLX_SGIX_pbuffer, GLX_SGIX_visual_select_group,
GLX_EXT_texture_from_pixmap, GLX_INTEL_swap_event
GLX version: 1.4
GLX extensions:
GLX_ARB_get_proc_address, GLX_ARB_multisample, GLX_EXT_import_context,
GLX_EXT_visual_info, GLX_EXT_visual_rating,
GLX_MESA_multithread_makecurrent, GLX_OML_swap_method,
GLX_SGI_make_current_read, GLX_SGIS_multisample, GLX_SGIX_fbconfig,
GLX_SGIX_pbuffer, GLX_SGIX_visual_select_group,
GLX_EXT_texture_from_pixmap
OpenGL vendor string: VMware, Inc.
OpenGL renderer string: Gallium 0.4 on llvmpipe (LLVM 0x301)
OpenGL version string: 2.1 Mesa 8.0.4
OpenGL shading language version string: 1.20
OpenGL extensions:
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_multitexture, GL_IBM_multimode_draw_arrays,
GL_IBM_texture_mirrored_repeat, 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_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_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_EXT_stencil_two_side, GL_EXT_texture_cube_map, GL_NV_fog_distance,
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_combine3, GL_ATI_texture_float, GL_EXT_shadow_funcs,
GL_EXT_stencil_wrap, GL_MESA_pack_invert, GL_MESA_ycbcr_texture,
GL_NV_primitive_restart, GL_ARB_fragment_program_shadow,
GL_ARB_half_float_pixel, GL_ARB_occlusion_query2, 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_pixel_buffer_object, GL_ARB_texture_compression_rgtc,
GL_ARB_texture_float, GL_ARB_texture_rectangle,
GL_ATI_texture_compression_3dc, GL_EXT_packed_float,
GL_EXT_pixel_buffer_object, GL_EXT_texture_compression_rgtc,
GL_EXT_texture_mirror_clamp, 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_ARB_vertex_array_object, GL_ATI_separate_stencil,
GL_ATI_texture_mirror_once, GL_EXT_draw_buffers2, GL_EXT_draw_instanced,
GL_EXT_gpu_program_parameters, GL_EXT_texture_compression_latc,
GL_EXT_texture_sRGB_decode, GL_OES_EGL_image, GL_ARB_copy_buffer,
GL_ARB_draw_instanced, GL_ARB_half_float_vertex, GL_ARB_instanced_arrays,
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_AMD_draw_buffers_blend,
GL_ARB_ES2_compatibility, GL_ARB_draw_buffers_blend,
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_shader_texture_lod,
GL_ARB_vertex_type_2_10_10_10_rev, GL_EXT_provoking_vertex,
GL_EXT_texture_snorm, GL_MESA_texture_signed_rgba, GL_ARB_robustness,
GL_ARB_texture_storage

First: I use a nVidia card with the nouveau driver. I only have one monitor to test on.

The console size is determined directly according to the chosen KMS mode. However, currently the screen buffer size is initialized statically to some default (80 cols x 24 rows). The font is scaled to match the screen size and the supposed size of a single cell in the console buffer.

This seems to be backwards to me. Since kmscon is a full screen app, what the user should pick is the font size, not the number of rows and columns. So, for example, instead of:
Mode: 1280 x 1024, Buffer 80 x 24 -> Font height 1024 / 24 = 42 (or something like that)
What should be calculated is:
Mode: 1280 x 1024, Font size (10pt) -> Buffer 160 x 70 (or something like that).

The only thing holding this back, as far as I can tell, is that the terminal actually supports multiple outputs. I can't test it, so I'm not sure how that is supposed to work. If I have two monitors, 1280x1024 and 800x600, how would you clone a single console to both of them in a reasonable way? Is that even expected to work?

I'll be happy to hear your thoughts.

Using the freetype2 backend the horizontal advance looks better, but the line-spread i so short that descenders are clipped (e.g. lower bowl of a g)

The latest release kmscon-6, which is packaged by Archlinux, does not have an important fix regarding multi-cell characters (commit 0de1de2, which happened after the release of kmscon-6). Could you do a new release so that Archlinux maintainer can package the fix?

After entering exit the terminal window doesn't close itself, I have to click on the leftmost window button.

Hi David,
first let me say that I don't know what I'm talking about :)

Currently the default font size is 15. It was changed from 10, which appeared on the screen extremely small and illegible. Now in most applications I use (e.g. X, firefox, openoffice...) 10-11 is the "normal" size and 15 is pretty large. Changing the KMSCON_FONT_DEFAULT_PPI to 96 makes the font sizes appear like what I would expect from other applications.

I don't really understand exactly what effect that has. I read the code comments and you probably chose the 72 value for a reason. Just thought I'd ask if that's the intended behavior.

Thanks!
Ran

Minimal version of FUSE is 2.9.0, but version checking is not performed.

If version of FUSE less than 2.9.0, build fails on src/kmscon_cdev.c because it uses "fuse_buf" struct on line 1214, but in this case (version of FUSE less than 2.9.0) there is no in fuse/fuse_common.h

It would be nice if kmscon supported simple screen blanking (with a configureable timeout).

I think the 8x16 font doesn’t have these characters

┬─>
└─>

Using the pango backend characters are smushed together and are set too narrow.

Is there a way to capture the output in an image?

Hi David,

Starting ./kmscon (current HEAD: 9d2018f) doesn't show anything, besides some strange residual text content from a previous run on the top of the screen (e.g. some "ran $ ls" prompt cut in the vertical middle, which is probably what was there before).

The machine on which this happens has an nvidia 9600GT with the nvidia driver (yea, sorry). kmscon uses --fbdev here. Here is the lspci for it:

01:00.0 VGA compatible controller: NVIDIA Corporation G94 [GeForce 9600 GT](rev a1) (prog-if 00 [VGA controller])
Flags: bus master, fast devsel, latency 0, IRQ 16
Memory at f6000000 (32-bit, non-prefetchable) [size=16M]
Memory at e0000000 (64-bit, prefetchable) [size=256M]
Memory at f4000000 (64-bit, non-prefetchable) [size=32M]
I/O ports at c000 [size=128]
[virtual] Expansion ROM at f7000000 [disabled] [size=512K]
Capabilities:
Kernel driver in use: nvidia

There are no other GPUs on this machine. Kernel version is 3.5.4.

Bisection:
5be8972 is the first bad commit
commit 5be8972
Author: David Herrmann [email protected]
Date: Sat Oct 27 18:00:27 2012 +0200

kmscon: add --primary-gpu-only and --all-gpus options

By default, kmscon now only uses primary und auxiliary displays. All
uncategorized displays are ignored. This fixes problems with dual-GPU
systems.

--primary-gpu-only makes kmscon not use any auxiliary displays. --all-gpus
makes kmscon also use uncategorized displays.

Signed-off-by: David Herrmann <[email protected]>

:040000 040000 2540782870b28321cccd9f468c36324c0001f45b 1a2cfa8fd88f08160c2fc5d3f6649efbbae4d4a0 M src

Passing --all-gpus makes it work normally.

Here is a --debug log:
https://gist.github.com/3965344

I wrote a quick patch to hook up maximize http://lists.freedesktop.org/archives/wayland-devel/2012-September/005587.html

Cheers

Running without --dumb I get a lot of debug spew.

sudo ./kmscon --xkb-layout de --xkb-variant nolatrash --font-engine pango --font-size 16 --font-name Pragmata

Then I switch to VT 1, there’s nothing to see, so I switch back and it outputted thousands of lines of warnings:

[0000.000000] NOTICE: kmscon Build #1 Aug 18 2012 17:40:12
[0000.024509] NOTICE: vt: using tty /dev/tty1
[0000.027251] INFO: new seat seat0
[0000.110814] INFO: video_drm: probing /dev/dri/card0
[0000.137614] INFO: video_drm: new drm device via /dev/dri/card0
[0000.140174] INFO: video: new device 0xb3ec90
[0001.114160] INFO: video: new display 0xb4e340
[0001.114236] INFO: video_drm: display 0xb4e340 DPMS is ON
[0001.114372] INFO: video_drm: activating display 0xb4e340 to 1440x900
[0001.142757] INFO: video_drm: setting DPMS of display 0xb4e340 to ON
[0001.253697] WARNING: text_gltex: cannot load glyph data into OpenGL texture (1280)
[0001.253758] WARNING: text_gltex: cannot load glyph data into OpenGL texture (1280)
[0001.253877] WARNING: text_gltex: cannot load glyph data into OpenGL texture (1280)
[0001.254010] WARNING: text_gltex: cannot load glyph data into OpenGL texture (1280)
[0001.254159] WARNING: text_gltex: cannot load glyph data into OpenGL texture (1280)
[0001.254307] WARNING: text_gltex: cannot load glyph data into OpenGL texture (1280)
[0001.254340] WARNING: text_gltex: cannot load glyph data into OpenGL texture (1280)
[0001.254478] WARNING: text_gltex: cannot load glyph data into OpenGL texture (1280)
[0001.254717] WARNING: text_gltex: cannot load glyph data into OpenGL texture (1280)
[0001.254752] WARNING: text_gltex: cannot load glyph data into OpenGL texture (1280)

The last line repeats 10000 times even though I was on VT 1 only for a second.

Kernel command line fbcon=rotate:3, or else # echo 3 > /sys/class/graphics/fbcon/rotate_all rotates the linux console.

I think it's necessary for kmscon to have rotation ability at some point.

I've just set up kmscon (7) under Arch Linux (kernel 3.7.9-2) on an IBM ThinkPad T40:

# cat /proc/fb
0 radeondrmfb

The first time kmscon launches it works correctly. If I switch to the kernel text console and then back to kmscon, the display is now corrupt (see this image).

After this point, launching any new kmscon instances will behave the same way.

This is a list of items that are still to do. If you absolutely want to see some feature or if you have special notes for a feature on this list, please open a separate bug/feature-report for this single TODO-item only.

kmscon:

• print /etc/issue on setup (try to be 100% compatible to agetty)
• per output configuration
• add input/video device blacklist
• add proper .ini-style configuration file
• make --gpus=primary fall back to other GPUs in single-gpu mode
• reload configuration on SIGHUP
• Avoid running multiple --listen instances on the same seat
• Allow different processes to create cdev sessions
• dynamic keyboard modifier detection
• fail if a terminal cannot be started (at least in non-listen mode) so kmscon exits
• read systemd-localed on startup to retrieve x11-keyboard layout
• make systemd units use $(prefix) instead of hard-coding
• install systemd units
• background images
• draw underline in text renderers
• rework horrible shl_ring implementation

tsm:

• support for VT220-style mouse-protocol
• "fill" state of lines and buffers in tsm-screen
• don't redraw buffer on keypress but on state-change
• Partial redraws
• multi-width characters (wcwidth())
• Support Unicode combining marks
• add c++ 'extern "C"' markers
• configurable palette
• "no bold" option

wlterm:

• add support for window-title OSM and hook it up with wlt-toolkit
• implement "snap-from-maximize" so you can "unmaximize" windows my moving them far enough

uterm:

• make uterm_video_wake_up() remove displays on failure
• handle SYN_DROPPED input-events
• cursor support
• delay device-probe if device changes seats (due to race-condition)
• make udev optional
• keep devices for unknown seats around until the seat shows up and then assign them
• figure out dead-key support
• make VT-switching keys configurable

On archlinux, I setup kmscon to replace agetty by creating a symlink in /etc/systemd/system/ between [email protected] and [email protected]. This works great as systemd now automatically starts kmscon instead of agetty.

Now typically I use startx in place on a tty. So when I launch xmonad through startx on tty1, it launches X on tty1 instead of creating allocating new tty. This is necessary to have startx work nicely with loginctl. If startx instead grabs a new tty, according to systemd, the login is no longer considered active.

So when running startx on tty1, I essentially run startx -- vt0

However it seems like running kmscon prevents access to the underlying virtual terminal. Running the above command yields the following error:

Fatal server error:
xf86OpenConsole: VT_ACTIVATE failed: No such device or address

Basically, afaik, X can't draw on top of kmscon like it could with agetty.

I don’t know if there’s a way to screenshot this, so here’s only a description to repro:

$ vim
:colo shine

kmscon doesn’t draw the background of fullscreen apps that set a non-back background color for the parts of the lines that have no foreground text. This is probably related to a reset of the background color that should not take place.

On the other hand I have this little script I wrote that works with background colors (The problem here is that cursor query returns a weird value, so it breaks): nyan.sh

On my environment shift+tab is not working in kmscon tty terminal.
In the normal gettty terminal during pressing "Ctrl+V Shifnt+Tab" in the terminal appreaed ^[[Z. In the kmscon nothing, even tab.

As I understarding in kmscon is using xkbd , not kbd as in normal tty. Any suggestion how to resolve this?

Thanks,
Vladimir

Not sure if that has anything to do with termcap, but I cannot use things like <Alt>Period or <Alt>Backspace

There no fealing that I working in the real console. Cursor in blinking =)

Will great if there will options related to cursor (box size, blinking refresh rate and so on)

Sometimes when I try to resize the terminal window it freezes and I can't enter anything except for closing the window.

label: bikeshed

It would be cool it I weren’t forced to use gray antialias. My fontconfig sets rgba rendering and LCD filter. Is it wrong to assume that Pango can handle this automatically unless forced? Elsewhere I always get LCD filtered text with Pango.

Otherwise, I would prefer the built-in bitmap font. So either RGBA antialias or black and white.

In bare kmscon (without tmux)
journalctrl -b ends gracefully after pressing q
journalctl after pressing q acts for a longer moment like cat > /dev/null and takes 100% of CPU. Then it exits without any error

$PAGER / $SYSTEMD_PAGER are empty
systemd version - 196

Happy bug-hunting season!

Cannot compile kmscon on Archlinux. Such error is appeared.

CC src/kmscon-kmscon_dummy.o
CC src/kmscon-kmscon_terminal.o
CC src/kmscon-kmscon_cdev.o
make[2]: *** No rule to make target libfont.la', needed bykmscon'. Stop.
make[2]: Leaving directory /tmp/packerbuild-0/kmscon-git/kmscon-git/src/kmscon-build' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory/tmp/packerbuild-0/kmscon-git/kmscon-git/src/kmscon-build'
make: *** [all] Error 2
==> ERROR: A failure occurred in build().
Aborting...
The build failed.

It would be nice if the font face/size/etc could be set dynamically inside a kmscon session. I certainly find myself in situations where my selected font size turns out to be a bad fit for some information (e.g., I want to display a really wide table and want to crank the font size down).

Could not input cjk without input mothod

Function keys F13 - F20 are under <Shift>F3 - <Shift>F10

Midnight Commander uses them https://github.com/MidnightCommander/mc/blob/master/lib/tty/key.c

On Ubuntu 12.04, Intel Core i5 video GPU, when I run -l option I get:
ERROR: cannot initialize kmscon, error: -14, Bad address
(or very similar ;)).

I'm using udev-171 and receive the following error when I'm trying to compile kmscon-6:

libtool: link: x86_64-pc-linux-gnu-gcc -std=gnu99 -Wall -O0 -march=native -O2 -pipe -Wl,-O1 -o .libs/kmscon external/kmscon-htable.o src/kmscon-conf.o src/kmscon-log.o src/kmscon-pty.o src/kmscon-text.o src/kmscon-kmscon_seat.o src/kmscon-kmscon_conf.o src/kmscon-kmscon_main.o src/kmscon-kmscon_dummy.o src/kmscon-kmscon_terminal.o src/kmscon-kmscon_cdev.o src/kmscon-text_bblit.o src/kmscon-text_bbulk.o src/kmscon-text_gltex.o src/kmscon-static_gl_math.o src/kmscon-static_gl_shader.o src/kmscon-static_shaders.o -pthread -Wl,--as-needed ./.libs/libeloop.so ./.libs/libuterm.so /usr/lib64/libudev.so -lrt -lpciaccess -ldrm -lEGL -lgbm ./.libs/libtext-font.a /var/tmp/portage/x11-terms/kmscon-6/work/kmscon-6/.libs/libtsm.so -lpangoft2-1.0 -lpango-1.0 /usr/lib64/libfontconfig.so -lfreetype -lz -lbz2 -lexpat -lgobject-2.0 -lglib-2.0 -lpthread ./.libs/libtsm.so /usr/lib64/libxkbcommon.so -lfuse -lGLESv2 -pthread
./.libs/libuterm.so: undefined reference to udev_device_has_tag' collect2: ld returned 1 exit status make[2]: *** [kmscon] Error 1 make[2]: Leaving directory/var/tmp/portage/x11-terms/kmscon-6/work/kmscon-6'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/var/tmp/portage/x11-terms/kmscon-6/work/kmscon-6'
make: *** [all] Error 2

The following configuration was used:

configure: Build configuration:

           prefix: /usr
      exec-prefix: ${prefix}
           libdir: /usr/lib64
       includedir: ${prefix}/include

Applications and Libraries:
kmscon: yes (yes: )
wlterm: no (no: enable-wlterm)
uterm: yes (yes: )
tsm: yes (yes: )
eloop: yes (yes: )

Miscellaneous Options:
debug: no (no: enable-debug)
optimizations: no (no: enable-optimizations)
multi-seat: no (no: enable-multi-seat)
hotplug: yes (yes: )
pciaccess: yes (yes: )
eloop-dbus: no (no: enable-eloop-dbus)

Video Backends:
fbdev: yes (yes: )
dumb drm: yes (yes: )
drm: yes (yes: )

Font Backends:
8x16: yes (yes: )
unifont: no (yes: )
freetype2: yes (yes: )
pango: yes (yes: )

Renderers:
bblit: yes (yes: )
bbulk: yes (yes: )
gltex: yes (yes: )

Session Types:
dummy: yes (yes: )
terminal: yes (yes: )
cdev: yes (yes: )

One of the widely used ones is $COLORTERM

It's needed for eg. automatically spawning instance of tmux (which i don't want to do in gnome-terminal/Terminal)

Couple of terminals that sets COLORTERM — gnome-terminal, Terminal, rxvt-xpm
One that doesn't — terminology

Most of those which are running under X11 are using xterm-256color terminfo/termcap, so there was a need to differentiate them from the shell script in this way ;)

This patch fixes a bunch of typos:
https://gist.github.com/jwilk/4751194

When running test_output I assumed that the output with --fbdev would be identical to the output without --fbdev. It is not.

http://i.imgur.com/rMB1t.jpg

There is some nasty bug in the TSM-screen layer. It causes a SIGSEGV in free() so there might be some random memory corruption (or double free that isn't properly detected). I cannot narrow it down but it happens when setting the "perfect" combination of a bottom-margin + reszizing often + cat /dev/urandom.

Just as reminder that we need to check that again. If someone finds a way to reliably reproduce it, I will try narrowing it down further.