]> git.ipfire.org Git - thirdparty/plymouth.git/log
thirdparty/plymouth.git
14 years agomain: fix kernel command line parsing
Ray Strode [Wed, 12 Oct 2011 19:50:56 +0000 (15:50 -0400)] 
main: fix kernel command line parsing

check_for_consoles has another bug (surprised?) where it would
jump too many characters forward if the command line has
console=tty0 in it, since tty0 is transparently changed to /dev/tty1.

14 years agomain: plug fd leak
James Hunt [Wed, 12 Oct 2011 02:11:10 +0000 (22:11 -0400)] 
main: plug fd leak

This plugs an fd leak in an error path

14 years agoutils: drop ply_list_directory
Ray Strode [Wed, 12 Oct 2011 02:09:37 +0000 (22:09 -0400)] 
utils: drop ply_list_directory

It's unused code

14 years agoutils: plug dir fd leak
James Hunt [Wed, 12 Oct 2011 02:08:48 +0000 (22:08 -0400)] 
utils: plug dir fd leak

This commit calls closedir() in an error path.

14 years agologger: plug fd leak
James Hunt [Wed, 12 Oct 2011 02:05:54 +0000 (22:05 -0400)] 
logger: plug fd leak

This plugs an fd leak in an error path.

14 years agotext-progress-bar: plug fd leak
James Hunt [Wed, 12 Oct 2011 02:03:08 +0000 (22:03 -0400)] 
text-progress-bar: plug fd leak

This plugs a fd leak in an error path

14 years agoclient: fix fd leak
James Hunt [Wed, 12 Oct 2011 02:01:07 +0000 (22:01 -0400)] 
client: fix fd leak

This plugs an fd leak in an error path

14 years agoREADME: resync from wiki
Ray Strode [Mon, 10 Oct 2011 14:54:00 +0000 (10:54 -0400)] 
README: resync from wiki

walters pointed out the README had some stale information in
it.  I've copy and pasted the wiki text in, now, so there
will be hopefully a small quantity of said stale information.

14 years agoautogen.sh: rework to be more buildsystem friendly
Colin Walters [Sun, 9 Oct 2011 19:54:15 +0000 (15:54 -0400)] 
autogen.sh: rework to be more buildsystem friendly

One of the goals laid out in
http://people.gnome.org/~walters/docs/build-api.txt
is to get projects to have autogen.sh scripts that
don't invoke configure after generating it.

This commit does that along with some other small
modifications designed to help be more confromant
with the Build API referenced in the above url.

https://bugs.freedesktop.org/show_bug.cgi?id=41629

14 years agoevent-loop: use static buffer instead of stack allocation buffer
James Hunt [Fri, 7 Oct 2011 15:32:46 +0000 (16:32 +0100)] 
event-loop: use static buffer instead of stack allocation buffer

alloca() isn't a good idea to use, in general, because it provides
undefined behavior when it fails.

This commit changes the epoll event buffer from being stack allocated
to static to avoid any potentially problems resulting from stack
space exhaustion.

https://bugs.freedesktop.org/show_bug.cgi?id=41562

14 years agoREADME: spelling fix.
Ville Skyttä [Sun, 25 Sep 2011 21:27:17 +0000 (00:27 +0300)] 
README: spelling fix.

This changes the mispelled word analagous to the
word analogous.

https://bugs.freedesktop.org/show_bug.cgi?id=41196

14 years agothrobgress: add plugin decelartion
Ray Strode [Fri, 19 Aug 2011 21:32:36 +0000 (17:32 -0400)] 
throbgress: add plugin decelartion

This should fix the build.

https://bugs.freedesktop.org/show_bug.cgi?id=40240

14 years agothrobgress: make colors configurable
Leho Kraav [Thu, 18 Aug 2011 15:06:17 +0000 (18:06 +0300)] 
throbgress: make colors configurable

This commit makes the background start and
background end colors configurable in throbgress.

14 years agoRevert "Add --nowait option."
Ray Strode [Wed, 3 Aug 2011 18:00:25 +0000 (14:00 -0400)] 
Revert "Add --nowait option."

This reverts commit c926bccd83ed1391df80adfecd1229b0327e1f11.

It was an in-progress patch that I carelessly and erroneously
pushed.

14 years agox11: translate enter to new line
Ray Strode [Fri, 29 Jul 2011 19:16:50 +0000 (15:16 -0400)] 
x11: translate enter to new line

commit 4081bd29fb471a93daba746643419d858f06bdfe changed
our keyboard handling code to treat newline instead of
carriage return as the enter key.

That commit fixed up the tty settings to work under these
new assumptions.  It failed to update the X11 code, though.

This commit fixes the X11 renderer plugin so it works in
the same way.

14 years agoAdd --nowait option.
cee1 [Wed, 13 Jul 2011 02:53:01 +0000 (10:53 +0800)] 
Add --nowait option.

With --nowait option, plymouth will not wait for reply from plymouthd.

14 years agoprogress-animation: allow 0 frame animations
cee1 [Sat, 18 Jun 2011 02:11:47 +0000 (10:11 +0800)] 
progress-animation: allow 0 frame animations

Currently ply-progress-animation crashes if you try to pass it
an animation with 0 available frames.

This commit makes it silently succeed.

14 years agotwo-step plugin: free plugin->animation_dir when destroy
cee1 [Wed, 15 Jun 2011 09:18:42 +0000 (17:18 +0800)] 
two-step plugin: free plugin->animation_dir when destroy

14 years agokey-file: minor code shuffle
Ray Strode [Tue, 14 Jun 2011 14:04:25 +0000 (10:04 -0400)] 
key-file: minor code shuffle

Move things around to avoid false positives in coverity.

14 years agomain: fix console=/dev/tty special case
Canek Peláez Valdés [Tue, 31 May 2011 15:39:43 +0000 (11:39 -0400)] 
main: fix console=/dev/tty special case

commit 1d09ad7d51e803e56b52c4e84dd558a7ac4d2e39 contained a mistake
where it checked for "tty" twice instead of "tty" once and "/dev/tty"
once.

This commit fixes that error.

14 years agomain: special case console=tty
Ray Strode [Wed, 25 May 2011 18:28:42 +0000 (14:28 -0400)] 
main: special case console=tty

Some people use console=tty in the same way they'd expect
console=tty0 to work.

This commit makes them behave the same way.

14 years agokeyboard: handle kernel terminal EIO better
Ray Strode [Wed, 11 May 2011 14:28:47 +0000 (10:28 -0400)] 
keyboard: handle kernel terminal EIO better

In commit 89096d735f78fab300248962030d6554f9894011 we added
reopen retries when the kernel returns EIO to ply-terminal.c.

This is because when the kernel is closing a tty down, that
tty is unavailable to userspace to reopen.

Unfortunately, that commit neglected to inform the ply-keyboard
part of the code when the terminal retry was successful.  The
upshot of this, is that if plymouthd needs to retry opening the
tty, then the splash screens lose control over the keyboard.

This commit changes how input notification is sent to the keyboard
handling code, so the tty disconnects are transparent.

14 years agomain: allow log output to be redirected
Ray Strode [Tue, 10 May 2011 19:42:29 +0000 (15:42 -0400)] 
main: allow log output to be redirected

Sometimes putting debug output on the screen has strange side
effects.  This commit adds a new kernel command line option,
e.g:

plymouth.debug=stream:/dev/null

that moves the output "out of the way"

14 years agomain: write pid file and daemonize last
cee1 [Thu, 21 Apr 2011 19:21:19 +0000 (15:21 -0400)] 
main: write pid file and daemonize last

These actions serve to notify plymouthd's readiness.
We should make sure they are run after plymouthd is actually
done starting up.

14 years agomain: write pid file even when not daemonizing
cee1 [Thu, 21 Apr 2011 19:18:13 +0000 (15:18 -0400)] 
main: write pid file even when not daemonizing

This commit moves the pid file writing code to main
from ply_create_daemon, so that it gets run even when we
plymouthd isn't daemonized.

This is more symmetrical, anyway, since unlinking of the pid
file is handled in main.

14 years agotext-progress-bar: Add support for /etc/os-release
cee1 [Mon, 18 Apr 2011 09:15:25 +0000 (17:15 +0800)] 
text-progress-bar: Add support for /etc/os-release

Systemd uses '/etc/os-release' as release file, see
http://0pointer.de/public/systemd-man/os-release.html

14 years agomain: force frame-buffer plugin at shutdown
Ray Strode [Thu, 31 Mar 2011 03:22:09 +0000 (23:22 -0400)] 
main: force frame-buffer plugin at shutdown

This way we hopefully won't get kicked out of KD_GRAPHICS
mode when we're SIGKILLed causing the splash to disappear
prematurely.

14 years agomain: fix ply_mode_t/ply_boot_splash_mode_t confusion
Ray Strode [Thu, 31 Mar 2011 03:21:00 +0000 (23:21 -0400)] 
main: fix ply_mode_t/ply_boot_splash_mode_t confusion

They're very similar enums and I was using the wrong one,
which caused a harmless compiler warning.

14 years agorenderer: provide way to override renderer plugin
Ray Strode [Thu, 31 Mar 2011 03:14:05 +0000 (23:14 -0400)] 
renderer: provide way to override renderer plugin

This will allow us to force e.g. frame-buffer even
if drm would work.

14 years agomain: create runtime directory if it doesn't exist already
Ray Strode [Thu, 31 Mar 2011 02:51:27 +0000 (22:51 -0400)] 
main: create runtime directory if it doesn't exist already

This makes sure there is a convenient place to put pid files.

14 years agogdm: don't do legacy gdm transition at shutdown
Ray Strode [Tue, 29 Mar 2011 17:06:28 +0000 (13:06 -0400)] 
gdm: don't do legacy gdm transition at shutdown

14 years agosystemd: change instrumentation points for systemd
Ray Strode [Tue, 29 Mar 2011 06:18:36 +0000 (02:18 -0400)] 
systemd: change instrumentation points for systemd

We don't want to enable console verbosity until the splash is up,
and we want to disable it as soon as we aren't logging the console
anymore.

14 years agoconfigure: s/text/test/
Ray Strode [Tue, 29 Mar 2011 05:04:31 +0000 (01:04 -0400)] 
configure: s/text/test/

Just a typo from the previous commit

14 years agosystemd: toggle systemd messages when toggling console redirection
Ray Strode [Tue, 29 Mar 2011 02:14:34 +0000 (22:14 -0400)] 
systemd: toggle systemd messages when toggling console redirection

Normally systemd is very mute about messages.  It's important,
though for it to be chatty when plymouth is running so we can
show verbose messages to the user when they hit escape.

This commit adds a new --enable-systemd-integration configure
flag which explicitly tells systemd when to print messages to
the console.

This may get dropped in the future in lieu of init script
changes doing this instead of plymouth directly.

14 years agomain: consolidate ply_terminal_session_detach calls
Ray Strode [Tue, 29 Mar 2011 02:36:39 +0000 (22:36 -0400)] 
main: consolidate ply_terminal_session_detach calls

It's called in several places through the code. This commit
moves it into a function and just calls that function instead.

14 years agoclient: small leak fix
Ray Strode [Sat, 26 Mar 2011 05:02:49 +0000 (01:02 -0400)] 
client: small leak fix

14 years agoutils: drop some bizarre looking code
Ray Strode [Sat, 26 Mar 2011 04:56:30 +0000 (00:56 -0400)] 
utils: drop some bizarre looking code

14 years agosplash: add new spinner theme
Ray Strode [Wed, 23 Mar 2011 16:06:39 +0000 (12:06 -0400)] 
splash: add new spinner theme

This is a very simple theme with a small spinner on a dark background.

14 years agomain: add new --kernel-command-line option
Ray Strode [Wed, 23 Mar 2011 16:04:56 +0000 (12:04 -0400)] 
main: add new --kernel-command-line option

It's really annoying when you're running plymouthd in
debug mode and don't happen to have "splash" on your kernel
command line and then need to reboot.

This commit adds a new debug option to override the kernel
command line for ths system.

14 years agoprogress-animation: fail load if there are no frames
Ray Strode [Wed, 23 Mar 2011 16:04:05 +0000 (12:04 -0400)] 
progress-animation: fail load if there are no frames

If there are no progress animation frames we need to
fail from the load function so plugins can deal with it
gracefully..

14 years agotwo-step: make progress animation optional
Ray Strode [Wed, 23 Mar 2011 15:32:14 +0000 (11:32 -0400)] 
two-step: make progress animation optional

14 years agoupstart-bridge: add new helper program
Colin Watson [Sat, 12 Mar 2011 21:31:15 +0000 (16:31 -0500)] 
upstart-bridge: add new helper program

This commit adds a new program, plymouth-upstart-bridge,
the listens for upstart state changes and sends them to plymouth,
or prints them out as appropriate.

14 years agoclient: allow reply handlers to be optional.
Colin Watson [Sat, 12 Mar 2011 21:31:15 +0000 (16:31 -0500)] 
client: allow reply handlers to be optional.

This commit lightens the hard requirement that every
client request has to have a function handling the reply.

There's really no reason to be so strict, and it makes
life easier for none /bin/plymouth clients this way.

14 years agodrm: don't light up unlit monitors
Ray Strode [Wed, 9 Mar 2011 22:22:03 +0000 (17:22 -0500)] 
drm: don't light up unlit monitors

If a monitor is dark when plymouth is started, we shouldn't
try to light it up.  There are rules on which outputs can be
attached to which controllers and we're very likely to break
those rules just assigning things willy-nilly.

The kernel should set us up in the way it thinks is best anyway.
There's no reason to second guess the kernel here.

14 years agoterminal: Keep retrying to reopen tty on EIO after hang up
Ray Strode [Fri, 4 Mar 2011 19:57:52 +0000 (14:57 -0500)] 
terminal: Keep retrying to reopen tty on EIO after hang up

If the tty hangs up on us, the kernel will return EIO while it's
closing down.  As far as I know, there's no way to get notified
when it's back up again, so just keep retrying in a loop over
the course of a second or so.  Eventually give up, though.

This is really suboptimal and suggests we should get away from
using ttys for input at some point.

14 years agoclient: do not wait forever if daemon is not started
Andrey Borzenkov [Thu, 3 Mar 2011 01:51:32 +0000 (20:51 -0500)] 
client: do not wait forever if daemon is not started

Currently plymouth --wait will wait forever if daemon is not
started. This does not look right - we were asked to wait for
daemon to stop and daemon is obviously stopped. So make it
exit right away in this case.

This avoids timeouts during systemd boot if plymouth happened to be
stopped (or never started) before plymouth-quit-wait had chance to
start.

Signed-off-by: Andrey Borzenkov <arvidjaar@gmail.com>
14 years agoterminal: apply terminal settings immediately
Ray Strode [Tue, 1 Mar 2011 22:26:16 +0000 (17:26 -0500)] 
terminal: apply terminal settings immediately

14 years agoutils: don't depend on PAGE_SIZE
Ray Strode [Mon, 28 Feb 2011 18:01:39 +0000 (13:01 -0500)] 
utils: don't depend on PAGE_SIZE

It's not defined for all architectures.

We don't really need the page size anyway, we just need a reasonably
large value.

14 years agomain: convert tab to spaces
Ray Strode [Fri, 25 Feb 2011 18:59:16 +0000 (13:59 -0500)] 
main: convert tab to spaces

14 years agotests: don't run during make check
Ray Strode [Sat, 19 Feb 2011 22:10:34 +0000 (17:10 -0500)] 
tests: don't run during make check

They aren't unit tests so running them in make check is wrong.

Eventually, I'd like them to be unit tests, though, and not
just little scratch programs.

14 years agoterminal: unlock tty before mucking with it
Ray Strode [Fri, 18 Feb 2011 18:45:55 +0000 (13:45 -0500)] 
terminal: unlock tty before mucking with it

It seems that sometimes when the tty is stolen from us
it remains locked when we get it back.  That would be okay
if the attributes were still kosher, but they aren't.

This commit unlocks the tty before trying to reinitialize
the terminal with suitable attributes.

Debugged by Andrey Borzenkov <arvidjaar@gmail.com>

14 years agoclient: clean up warnings
Ray Strode [Wed, 9 Feb 2011 20:28:00 +0000 (15:28 -0500)] 
client: clean up warnings

14 years agodrm: clean up warnings
Ray Strode [Wed, 9 Feb 2011 20:26:59 +0000 (15:26 -0500)] 
drm: clean up warnings

14 years agoframe-buffer: clean up warnings
Ray Strode [Wed, 9 Feb 2011 20:26:21 +0000 (15:26 -0500)] 
frame-buffer: clean up warnings

14 years agotwo-step: fix warnings
Ray Strode [Wed, 9 Feb 2011 20:07:03 +0000 (15:07 -0500)] 
two-step: fix warnings

14 years agospace-flares: clean up warning
Ray Strode [Wed, 9 Feb 2011 20:03:43 +0000 (15:03 -0500)] 
space-flares: clean up warning

14 years agotext: clean up warning
Ray Strode [Wed, 9 Feb 2011 20:00:31 +0000 (15:00 -0500)] 
text: clean up warning

14 years agofade-throbber: clean up warnings
Ray Strode [Wed, 9 Feb 2011 19:59:45 +0000 (14:59 -0500)] 
fade-throbber: clean up warnings

14 years agothrobgress: clean up warnings
Ray Strode [Wed, 9 Feb 2011 19:58:33 +0000 (14:58 -0500)] 
throbgress: clean up warnings

14 years agolabel: clean up warnings
Ray Strode [Wed, 9 Feb 2011 19:56:58 +0000 (14:56 -0500)] 
label: clean up warnings

14 years agomain: detach from terminal session on failure
Ray Strode [Wed, 9 Feb 2011 19:54:06 +0000 (14:54 -0500)] 
main: detach from terminal session on failure

14 years agomain: make duplicate plymouthd's a silent error
Ray Strode [Wed, 9 Feb 2011 19:47:54 +0000 (14:47 -0500)] 
main: make duplicate plymouthd's a silent error

14 years agomain: clean up warnings
Ray Strode [Wed, 9 Feb 2011 19:44:56 +0000 (14:44 -0500)] 
main: clean up warnings

14 years agoimage: clean up warning
Ray Strode [Wed, 9 Feb 2011 19:40:20 +0000 (14:40 -0500)] 
image: clean up warning

14 years agotext-progress-bar: clean up warnings
Ray Strode [Wed, 9 Feb 2011 19:39:10 +0000 (14:39 -0500)] 
text-progress-bar: clean up warnings

14 years agoregion: clean up some warnings
Ray Strode [Wed, 9 Feb 2011 19:34:52 +0000 (14:34 -0500)] 
region: clean up some warnings

14 years agoarray: clean up a warning
Ray Strode [Wed, 9 Feb 2011 19:32:43 +0000 (14:32 -0500)] 
array: clean up a warning

14 years agoevent-loop: clean up some warnings
Ray Strode [Wed, 9 Feb 2011 19:31:01 +0000 (14:31 -0500)] 
event-loop: clean up some warnings

14 years agoanimation: add debug spew
Ray Strode [Wed, 9 Feb 2011 19:20:53 +0000 (14:20 -0500)] 
animation: add debug spew

14 years agotwo-step: Add more debug spew
Ray Strode [Wed, 9 Feb 2011 19:13:21 +0000 (14:13 -0500)] 
two-step: Add more debug spew

14 years agoboot-server: improve debug output
Ray Strode [Thu, 3 Feb 2011 20:28:05 +0000 (15:28 -0500)] 
boot-server: improve debug output

We've written "could not write bytes" in many different
places.  This changes the error message to be more domain
specific.

14 years agoboot-server: fix debug output when message is from init
Ray Strode [Thu, 3 Feb 2011 19:40:43 +0000 (14:40 -0500)] 
boot-server: fix debug output when message is from init

We were always trying to show the parent process name,
but init doesn't have a parent.

This commit special cases init, since it's a special case.

14 years agodrm: reduce minimum build requirements
Lucian Muresan [Tue, 1 Feb 2011 19:05:45 +0000 (14:05 -0500)] 
drm: reduce minimum build requirements

This patch adds the respective configure options to make it possible to
disable libdrm_intel, libdrm_radeon, libdrm_nouveau, and libkms
independently from each other.

https://bugs.freedesktop.org/show_bug.cgi?id=29804

14 years agopixel-buffer: Fix overflow in blend_two_pixel_values
Martin Pitt [Fri, 14 Jan 2011 23:32:10 +0000 (18:32 -0500)] 
pixel-buffer: Fix overflow in blend_two_pixel_values

It was using 16bit types to hold the intermiediate
results of the blend, but for high intensity,
low opacity input values it could need up to
17 bits to prevent overflow.

https://bugs.freedesktop.org/show_bug.cgi?id=33129

14 years agomain: remove tty0 from list of fallback ttys
Ray Strode [Sat, 20 Nov 2010 00:35:52 +0000 (19:35 -0500)] 
main: remove tty0 from list of fallback ttys

It's not a valid fallback tty since writing to
it would create a feedback loop (since it's
redirected with /dev/console)

14 years agomain: Always translate tty0 to tty1
Ray Strode [Sat, 20 Nov 2010 00:33:34 +0000 (19:33 -0500)] 
main: Always translate tty0 to tty1

Another bug in check_for_consoles...

We can't ever write to tty0 directly, because
it is redirected (just like /dev/console).  Previously
we would translate the call to tty1, but commit

c40fd792b6edf931a6bbe2ec23518b57483a4e2f

broke that.  This commit fixes it again.

14 years agoterminal: don't rely on strlen(bytes) for write size
Ray Strode [Fri, 19 Nov 2010 20:27:12 +0000 (15:27 -0500)] 
terminal: don't rely on strlen(bytes) for write size

We're printing stuff to the terminal.  This may include
NUL bytes once in a while.  We shouldn't rely on strlen()
to determine how many bytes to write.

14 years agoboot-client: fix fallback socket support
Andrey Borzenkov [Thu, 18 Nov 2010 18:35:11 +0000 (21:35 +0300)] 
boot-client: fix fallback socket support

Commit 3ec007a4820e53aed6713620173bd997957754ec did not
properly check for success when connecting to the fallback
socket path and instead always failed.

This commit fixes the code to properly check for success and
proceed.

Signed-off-by: Andrey Borzenkov <arvidjaar@gmail.com>
14 years agoutils: if concrete/abstract socket selection
Andrey Borzenkov [Thu, 18 Nov 2010 18:35:11 +0000 (21:35 +0300)] 
utils: if concrete/abstract socket selection

Commit 9de731ed29db4670c1641aedd945c56fd1a3feed caused
plymouth to erroneously drop the leading '\0' on the old
abstract socket making fallback socket support not work.
As result new clients (after update) can no longer communicate
with old running daemons (such as from the initrd).

This commit restores the leading '\0' for both
abstract sockets, trimmed and non-trimmed.

Commit 3ec007a4820e53aed6713620173bd997957754ec did not
check for success with old socket path and failed
always. Properly check for success and proceed.

Signed-off-by: Andrey Borzenkov <arvidjaar@gmail.com>
14 years agomain: don't pause on crash
Ray Strode [Mon, 15 Nov 2010 19:01:50 +0000 (14:01 -0500)] 
main: don't pause on crash

plymouthd would previously pause() when crashing if debug
mode was enabled so that it could be attached to with a debugger.

pausing indefinitely during boot up is often a bad idea though.

This commit changes it to sleep 30 seconds instead.

14 years agoterminal: Don't set ISTRIP in terminal attributes
Ray Strode [Mon, 15 Nov 2010 18:56:01 +0000 (13:56 -0500)] 
terminal: Don't set ISTRIP in terminal attributes

In certain error paths plymouth would erroneously set
ISTRIP on the terminal.  This is bogus in the same way
the changes fixed by commit ea394383c54dda771dcf59e7ec11280fa945280b
were bogus.

14 years agoconfigure: move with_gdm_autostart_file=no to else clause
Ray Strode [Tue, 2 Nov 2010 20:05:39 +0000 (16:05 -0400)] 
configure: move with_gdm_autostart_file=no to else clause

doing it unconditionally seems to confuse autogoo.

14 years agoviewer: always dist desktop file
Ray Strode [Tue, 2 Nov 2010 19:23:06 +0000 (15:23 -0400)] 
viewer: always dist desktop file

Even if we aren't installing it, we still want it to get shipped
when doing "make dist"

15 years agoboot-server: don't print error when client goes away
Ray Strode [Wed, 20 Oct 2010 14:04:24 +0000 (10:04 -0400)] 
boot-server: don't print error when client goes away

Right now, anytime the server is unable to respond to a
client it puts an ugly message on the screen:

"could not write bytes: Broken pipe"

or some such.  That message isn't really useful unless you're
debugging your distribution, so change it from a ply_error to
ply_trace.

15 years agopopulate-initrd: pre-expand logofile variable
Ray Strode [Wed, 13 Oct 2010 12:26:42 +0000 (08:26 -0400)] 
populate-initrd: pre-expand logofile variable

The logofile by default is $datadir/plymouth.png

$datadir contains a reference to $datarootdir, so
we need to preexpand the variable in configure, for
the right value to get written to plymouth-populate-initrd.

15 years agopopulate-initrd: don't hardcode client and daemon path
Brett Witherspoon [Tue, 12 Oct 2010 03:05:24 +0000 (22:05 -0500)] 
populate-initrd: don't hardcode client and daemon path

When not installing in system root, the populate initrd script would not
install the client or daemon due to the paths being hardcoded.
Environmental variables are also now available to override the defaults.

15 years agoimage: replace deprecated libpng function
Brett Witherspoon [Tue, 12 Oct 2010 03:23:47 +0000 (22:23 -0500)] 
image: replace deprecated libpng function

The png_set_gray_1_2_4_to_8 function is deprecated and has been removed
from libpng14. Now png_set_expand_gray_1_2_4_to_8 is used instead which
is compatible with libpng-1.2.9 and greater.

15 years agoviewer: don't build viewer by default
Ray Strode [Mon, 11 Oct 2010 14:33:37 +0000 (10:33 -0400)] 
viewer: don't build viewer by default

The viewer is useful for seeing boot messages after boot up.
It does this by showing a notification icon in the event there
is a problem during boot.

Notification icons aren't as en vogue as they once were, however.

Ideally, we would have a more structured and semantically aware
way to deal with specific boot problems.

This commit turns the icon off by default.  It can still be built
with a --with-log-viewer

Bug: https://bugs.freedesktop.org/show_bug.cgi?id=30724
Reported By: William Jon McCann <william.jon.mccann@gmail.com>

15 years agoboot-server: make update handling asyncrhonous
Ray Strode [Thu, 7 Oct 2010 19:34:08 +0000 (15:34 -0400)] 
boot-server: make update handling asyncrhonous

This call needs to be cheap since some distros
do a ton of updates.

15 years agoutils: use system default max for listen()
Ray Strode [Thu, 7 Oct 2010 19:08:47 +0000 (15:08 -0400)] 
utils: use system default max for listen()

There's apparently a define that gives the right value to
pass to listen.  So this commit makes us use that instead of
our arbitrary "32"

15 years agolist: drop node_set_data method
Ray Strode [Thu, 7 Oct 2010 16:58:19 +0000 (12:58 -0400)] 
list: drop node_set_data method

It's not used anywhere, so drop it.

15 years agoprogress: plug small memory leak
Ray Strode [Thu, 7 Oct 2010 17:18:26 +0000 (13:18 -0400)] 
progress: plug small memory leak

We weren't freeing the message structures.

15 years agomain: NULL initialize "console" variable
Ray Strode [Thu, 7 Oct 2010 16:47:47 +0000 (12:47 -0400)] 
main: NULL initialize "console" variable

check_for_consoles is a really small function, that has historically
been loaded with bugs.

This commit fixes another bug in it, where the consoles variable is
never initialized to NULL.

15 years agobranch-merge: clean up socket name
Ray Strode [Wed, 6 Oct 2010 22:07:08 +0000 (18:07 -0400)] 
branch-merge: clean up socket name

The plymouth client talks to the plymouth daemon via
an abstract socket.  There aren't a lot of "rules" for
the name of the abstract socket, and so plymouth has just
used "/ply-boot-protocol" for the name.  While this is
perfectly valid, a number programs (hal, udev, etc) use
the convention "/org/freedesktop/program".

"/org/freedesktop/plymouthd" is certainly much nicer than
"/ply-boot-protocol" so we're going to change to that, and
fall back to the old name for compatibility.

One other niggle is trailing zeros in the name.  The socket
address is stored in a fixed size buffer.  Traditionally,
programms would pass the size of the entire socket structure,
including the full size of the aaddress buffer to bind and
connect.  This means that any NUL bytes in the address buffer
after the address become part of the address.  This means users
looking at /proc/net/unix will see all the extra NUL bytes.

One trick that some programs employee to skirt around this problem,
is to pass only the size of the structure less the trailing
NUL bytes of the address buffer to bind and connected.  While maybe
not 100% kosher, this works okay in practice.

plymouth will now use that trick as well.

15 years agoprotocol: change socket path
Ray Strode [Wed, 6 Oct 2010 22:03:25 +0000 (18:03 -0400)] 
protocol: change socket path

Before we were using

  /ply-boot-protocol\0\0\0\0...\0

for our address which really uglifies /proc/net/unix
and doesn't match what a lot of other programs do.

This commit changes the address to just

  /org/freedesktop/plymouthd

which is much nicer.

15 years agoprotocol: drop leading \0 from socket path
Ray Strode [Wed, 6 Oct 2010 21:57:34 +0000 (17:57 -0400)] 
protocol: drop leading \0 from socket path

The socket path is currently defined to be:

 #define PLY_BOOT_PROTOCOL_SOCKET_PATH "\0/ply-boot-protocol"

The \0 is because it's an abstract socket, and abstract sockets
have a leading NUL.  The code always ignores the NUL though and
adds it back later, so it's not needed.

This commit just drops it.

15 years agoutils: add support for "trimmed abstract" sockets
Ray Strode [Wed, 6 Oct 2010 21:43:50 +0000 (17:43 -0400)] 
utils: add support for "trimmed abstract" sockets

Right now plymouth listens on an abstract socket with a name
that has a bunch of trailing NUL bytes.  These zeroes uglify
the output of /proc/net/unix among other things.

This commit adds support for a new "trimmed abstract" socket
type, which drops the zeroes.

A subsequent commit will actually change plymouthd to use the
new api.

15 years agoset-default-theme: redirect to /dev/null not /null
Ray Strode [Mon, 4 Oct 2010 13:47:24 +0000 (09:47 -0400)] 
set-default-theme: redirect to /dev/null not /null

The set-default-theme script was incorrectly redirecting
stderr from a grep command to /null instead of /dev/null.

This causes a /null file to get created on the filesystem.

Bug: https://bugs.freedesktop.org/show_bug.cgi?id=30576
Reported by: David LeBlanc <leblancdw@yahoo.com>

15 years agoviewer: Add a title to the status icon
Ray Strode [Thu, 30 Sep 2010 03:08:45 +0000 (23:08 -0400)] 
viewer: Add a title to the status icon

15 years agoset-default-theme: exit 0 at bottom of file
Ray Strode [Wed, 29 Sep 2010 17:25:22 +0000 (13:25 -0400)] 
set-default-theme: exit 0 at bottom of file

This is so that if there is no --rebuild-initrd it still
exits successfully.