]> git.ipfire.org Git - thirdparty/plymouth.git/log
thirdparty/plymouth.git
16 years ago[script] Dereference on assign without affecting refcount
Charlie Brej [Mon, 22 Jun 2009 15:37:58 +0000 (16:37 +0100)] 
[script] Dereference on assign without affecting refcount

Assigns were incorrectly dereferencing creating an incorrect refcount.

16 years ago[script] Add support for logical AND and OR
Charlie Brej [Thu, 18 Jun 2009 16:27:32 +0000 (17:27 +0100)] 
[script] Add support for logical AND and OR

Allows the use of "&&" and "||". These are evaluated lazily and return the
evaluated sub-value which completed the operation rather than a bool. It allows
things like:

reply = cache_lookup(index) || slow_lookup(index);

If cache_lookup returns a false value (NULL, or 0) then slow_lookup is executed
and its result is placed in reply. Otherwise the result from cache_lookup is
used.

16 years ago[script] Add support for floats
Charlie Brej [Tue, 16 Jun 2009 15:33:29 +0000 (16:33 +0100)] 
[script] Add support for floats

Adds parsing and execution support for floats. Some operations upgrade ints to
floats with no way of turning them back to ints.

16 years agoInitial scripted plugin support
Charlie Brej [Fri, 12 Jun 2009 17:04:25 +0000 (18:04 +0100)] 
Initial scripted plugin support

This is an initial support for the scripted plugin. There are _many_ FIXMEs
and the whole code is reather unstable. The formatting is completely incorrect
and will be changed soon. There are scripts which are converted using a perl
to an C embeddable string.

16 years ago[frame-buffer] Discard overlapping flush areas
Ray Strode [Fri, 12 Jun 2009 03:11:26 +0000 (23:11 -0400)] 
[frame-buffer] Discard overlapping flush areas

One advantage of the previous bounding box approach to flushing
is overlapping flush areas wouldn't get flushed multiple times.

This commit tries to identify overlapping flush areas and eliminate
them. It's not perfect though.

A better approach might be to store a sorted tree of areas to be
flushed, and walk the tree when adding new flush areas to quickly find
overlapping areas. Then we'd split each area into two or more new areas
to avoid overlaps.

16 years ago[frame-buffer] Track multiple flush areas
Ray Strode [Fri, 12 Jun 2009 02:50:07 +0000 (22:50 -0400)] 
[frame-buffer] Track multiple flush areas

Previously we would always aggregate flush areas together even
if they were disjoint and far apart.  That meant that if two
images on opposites sides of the screen were updated in one
frame, then the entire screen would get redrawn.

We now track flush areas in a list, instead of a bounding box.

16 years ago[frame-buffer] Rework flush interface to take area
Ray Strode [Fri, 12 Jun 2009 02:31:11 +0000 (22:31 -0400)] 
[frame-buffer] Rework flush interface to take area

Before this commit it would always flush the area_to_flush
area. Now the interface allows flushing arbitrary areas. This
change paves the way for us to flush multiple disjoint areas
at one time.

16 years ago[splash] Add theme source files to EXTRA_DIST
Ray Strode [Wed, 6 May 2009 14:02:40 +0000 (10:02 -0400)] 
[splash] Add theme source files to EXTRA_DIST

They weren't making it into the dist'd tarball

16 years ago[solar] pull trigger right away if idle
Charlie Brej [Wed, 13 May 2009 16:08:35 +0000 (17:08 +0100)] 
[solar] pull trigger right away if idle

The idle_trigger was not being pulled in the solar plugin which causes the
system to deadlock on quit and plymouthd never exits.

16 years ago[two-step] pull trigger right away if already idle
Ray Strode [Mon, 4 May 2009 21:17:36 +0000 (17:17 -0400)] 
[two-step] pull trigger right away if already idle

The trigger always needs to get pulled so that the
daemon knows that it can quit.  Previously, we weren't
pulling it if two-step was already idle.  This meant that
plymouthd never quit.

16 years ago[two-step] Clean up animation when stopping
Ray Strode [Mon, 4 May 2009 04:09:34 +0000 (00:09 -0400)] 
[two-step] Clean up animation when stopping

16 years ago[two-step] Always finish 2nd step before quitting
Ray Strode [Mon, 4 May 2009 01:32:08 +0000 (21:32 -0400)] 
[two-step] Always finish 2nd step before quitting

Previously, we'd cut short in the middle of a very short animation.

16 years ago[daemon] Tell splash to become idle before quitting
Ray Strode [Mon, 4 May 2009 00:55:16 +0000 (20:55 -0400)] 
[daemon] Tell splash to become idle before quitting

This allows it to, e.g., finish up its current round of animations or
fade out the progress bar or whatever.

16 years ago[two-step] Add merge-fade as a progress animation transition type
Charlie Brej [Wed, 13 May 2009 12:11:36 +0000 (13:11 +0100)] 
[two-step] Add merge-fade as a progress animation transition type

The merge fade merges the two frames before drawing them to the framebuffer.
This makes a smooth transition between areas with partly transparent pixels or
areas in later frames becoming trasparent. This is set as the default on the
"glow" theme.

16 years ago[glow] Add fade between frames in glow theme
Ray Strode [Sat, 2 May 2009 05:03:41 +0000 (01:03 -0400)] 
[glow] Add fade between frames in glow theme

16 years ago[two-step] Add fade to progress animation
Ray Strode [Sat, 2 May 2009 04:36:28 +0000 (00:36 -0400)] 
[two-step] Add fade to progress animation

One issue with the two-step plugin is that during the initial boot
progress the frames can be updated several seconds apart from each
other.  This can look jumpy.  We can now optionally fade between
the frames to make the experience be a little smoother.

16 years ago[two-step] Add support for a "corner image"
Ray Strode [Sat, 2 May 2009 03:39:27 +0000 (23:39 -0400)] 
[two-step] Add support for a "corner image"

This is so the OLPC theme can have a
"Powered by blah" logo in the corner of the screen.

16 years ago[two-step] Pull background color from theme
Ray Strode [Sat, 2 May 2009 03:08:09 +0000 (23:08 -0400)] 
[two-step] Pull background color from theme

The OLPC theme looks best when it's white, so we need to be able to
specify the background color from the theme.  If the theme doesn't
specify a background color, we fall back to the old behavior of using
the distro default.

16 years ago[splash] Rename fade-in plugin to fade-throbber
Ray Strode [Thu, 14 May 2009 16:47:49 +0000 (12:47 -0400)] 
[splash] Rename fade-in plugin to fade-throbber

We're trying to keep a distinction between
plugin and theme.

16 years ago[splash] Rename solar plugin to space-flares
Ray Strode [Thu, 14 May 2009 14:10:27 +0000 (10:10 -0400)] 
[splash] Rename solar plugin to space-flares

We're trying to keep a distinction between
plugin and theme.

16 years ago[splash] Rename spinfinity plugin to throbgress
Ray Strode [Thu, 14 May 2009 14:10:27 +0000 (10:10 -0400)] 
[splash] Rename spinfinity plugin to throbgress

It's the theme that makes it spinfinity,
not the plugin.  The plugin is more generic.

16 years ago[splash] Rename glow plugin to two-step
Ray Strode [Fri, 1 May 2009 20:57:27 +0000 (16:57 -0400)] 
[splash] Rename glow plugin to two-step

Now the "glow" theme is just one possible manifestation of the
"two-step" plugin.

16 years ago[splash] Drop pulser plugin
Ray Strode [Thu, 14 May 2009 14:29:32 +0000 (10:29 -0400)] 
[splash] Drop pulser plugin

It's ugly and doesn't pass the quality sniff test.

16 years ago[splash] Split splashes into plugins and themes
Ray Strode [Wed, 29 Apr 2009 18:25:02 +0000 (14:25 -0400)] 
[splash] Split splashes into plugins and themes

Some of the plugins (well, the glow plugin) would be a lot more
versatile if they could be reused for multiple splashes with different
images.

This commit splits boot splashes into two parts, the plugin engine which
does all the dirty work, and the theme which says which plugin to use
and optionally how the plugin should work (using plugin specific
key/value pairs)

16 years ago[key-file] Use hashtables instead of lists
Charlie Brej [Wed, 13 May 2009 10:54:30 +0000 (11:54 +0100)] 
[key-file] Use hashtables instead of lists

Not really any performance advantage but it makes the code a little simpler.

16 years ago[libply] Add new keyfile class
Ray Strode [Thu, 30 Apr 2009 03:23:12 +0000 (23:23 -0400)] 
[libply] Add new keyfile class

This is a very lame keyfile parser.  We'll be able to use it to load
some metadata about the plugins.

The idea is that plugins will shift to being engines that can be
leveraged from multiple splash themes.  This way, for instance, the glow
plugin (which will be renamed) can be used to implement several
splashes.

16 years ago[window] Don't force text mode if debugging is on
Ray Strode [Tue, 5 May 2009 17:48:05 +0000 (13:48 -0400)] 
[window] Don't force text mode if debugging is on

It's not that much more work to hit ctrl-t, and now
that we log messages to plymouth-crash.log it's a lot
less useful to see them on screen.

16 years ago[daemon] Update tests to include compile correctly
Charlie Brej [Wed, 13 May 2009 09:31:44 +0000 (10:31 +0100)] 
[daemon] Update tests to include compile correctly

Adds ply-list.c/h to all tests automake files which use ply-logger.c.
Adds PLYMOUTH_TIME_DIRECTORY/PLUGIN/THEME_PATH when ply-boot-splash.c is used.
Adds PLY_BOOT_SPLASH_MODE_BOOT_UP to the ply_boot_splash_show call.

16 years ago[daemon] Free the debug buffer only after printing the last trace message
Charlie Brej [Wed, 13 May 2009 17:03:43 +0000 (18:03 +0100)] 
[daemon] Free the debug buffer only after printing the last trace message

Was segfaulting or printing failed assertios just before exitting.

16 years ago[daemon] Dump log to plymouth-crash.log on crash
Ray Strode [Tue, 5 May 2009 14:59:57 +0000 (10:59 -0400)] 
[daemon] Dump log to plymouth-crash.log on crash

This will be useful for debug purposes.

It's implemented by adding the concept of
"filters" to the logger code.  Each time a message
is logged it is passed through all registered
filter handlers and then the output of the filters
are what is actually logged.

In this case, we pass the messages through verbatim,
but also redirect them to a debug buffer.

We only do this redirection when plymouth:debug is
used.

16 years ago[throbber] Fix call to remove_frames
Ray Strode [Mon, 4 May 2009 01:22:44 +0000 (21:22 -0400)] 
[throbber] Fix call to remove_frames

ply-throbber was passing in the frames directly
instead of the throbber object

16 years ago[animation] Fix call to remove_frames
Ray Strode [Mon, 4 May 2009 01:22:44 +0000 (21:22 -0400)] 
[animation] Fix call to remove_frames

ply-animation was passing in the frames directly
instead of the animation object

16 years ago[progress-animation] Fix call to remove_frames
Ray Strode [Sat, 2 May 2009 04:34:52 +0000 (00:34 -0400)] 
[progress-animation] Fix call to remove_frames

ply-progress-animation was passing in the frames directly instead of the
progress object

16 years ago[event-loop] Don't crash by running removed timeouts
Ray Strode [Tue, 5 May 2009 15:57:21 +0000 (11:57 -0400)] 
[event-loop] Don't crash by running removed timeouts

We used to run and remove timeouts in a loop.  This
breaks the case where the dispatched timeout handler
removes the timeout itself.

We fix it by making timeout handling a two pass thing.
First we find which handlers need to be dispatched,
and move them to their own list.  Then we run through
the new list and dispatch the handlers.

16 years agoSwitch to tty 63 on shutdown
Ray Strode [Fri, 8 May 2009 03:43:46 +0000 (23:43 -0400)] 
Switch to tty 63 on shutdown

This is less than ideal, but side-steps weird interactions
between init, X, and plymouth.

16 years agoHang around until killed by init at shutdown
Ray Strode [Thu, 7 May 2009 18:46:03 +0000 (14:46 -0400)] 
Hang around until killed by init at shutdown

init does two rounds of killing at shutdown:
a round of SIGTERMs and a round of SIGKILLs.

We want to stay alive until the SIGKILL round
so we ignore the SIGTERM signal.

16 years agoPass the mode plymouthd is running to splashes
Ray Strode [Thu, 7 May 2009 18:38:50 +0000 (14:38 -0400)] 
Pass the mode plymouthd is running to splashes

We don't want to show progress bars and things for
shutdown, so pass the mode to the plugins so they can
just do a static image.

16 years agoAdd stub INSTALL file
Ray Strode [Sat, 2 May 2009 05:10:19 +0000 (01:10 -0400)] 
Add stub INSTALL file

16 years agoPause progress by default on password and question queries
Charlie Brej [Tue, 31 Mar 2009 21:13:20 +0000 (22:13 +0100)] 
Pause progress by default on password and question queries

Plymouth client ask-for-password and ask-question commands have their option
"--pause-progress" changed to "--dont-pause-progress". By default progress is
paused and "--dont-pause-progress" has to be supplied to not pause.

16 years agoMake sure to clean up ptm when detaching
Ray Strode [Thu, 19 Mar 2009 13:37:28 +0000 (09:37 -0400)] 
Make sure to clean up ptm when detaching

Otherwise we hit an assert next time we attach

16 years agodon't create a new splash every time show is called
William Jon McCann [Wed, 18 Mar 2009 22:54:56 +0000 (18:54 -0400)] 
don't create a new splash every time show is called

Reuse the existing splash.

16 years agoFix build when ENABLE_FADE_OUT is defined in spinfinity
Jonathan Greig [Mon, 16 Mar 2009 15:00:31 +0000 (11:00 -0400)] 
Fix build when ENABLE_FADE_OUT is defined in spinfinity

The variable "i" wasn't be declared, but was being used to
loop.

16 years agoSupport "splash" in addition to "rhgb" on kernel command line
Ray Strode [Fri, 13 Mar 2009 14:08:15 +0000 (10:08 -0400)] 
Support "splash" in addition to "rhgb" on kernel command line

We used "rhgb" because that's what Fedora users were used to
instinctively removing when graphical boot messes up.  This
isn't a very intuitive (or very upstream-y) name though.

Now we check for "rhgb" or "splash"

16 years agoWhen setting term to buffered without an attributes_saved, use generated ones.
Charlie Brej [Wed, 11 Mar 2009 11:04:49 +0000 (11:04 +0000)] 
When setting term to buffered without an attributes_saved, use generated ones.

A local term_attributes was constructed but not used.

16 years agoInitialise state->showing_details at show_splash
Charlie Brej [Wed, 11 Mar 2009 10:53:07 +0000 (10:53 +0000)] 
Initialise state->showing_details at show_splash

Only effects us when running without rhgb in the kernel line, by having to
press Esc twice to get the default splash.

16 years agoReset to text mode anytime we're about to print text
Ray Strode [Tue, 10 Mar 2009 18:46:59 +0000 (14:46 -0400)] 
Reset to text mode anytime we're about to print text

If we don't then someone during boot up may change
out from raw when we aren't looking.

Maybe we should use TIOCSLCKTRMIOS to prevent that
from happening? Or maybe we should bypass the tty
for input entirely and use evdev?

16 years agoRemove event handlers on destroy in plugins
Ray Strode [Mon, 9 Mar 2009 16:18:56 +0000 (12:18 -0400)] 
Remove event handlers on destroy in plugins

This is because we can end up in scenarios
where a plugin gets destroyed, and readded
without getting hidden in between.

16 years agoFactor handler registration code out into separate functions in plugins
Ray Strode [Mon, 9 Mar 2009 16:16:10 +0000 (12:16 -0400)] 
Factor handler registration code out into separate functions in plugins

This will make it easier to unregister the handlers from both
hide and destroy

16 years agoFix password entry dialogue in fade-in plugin.
Charlie Brej [Mon, 9 Mar 2009 14:51:08 +0000 (14:51 +0000)] 
Fix password entry dialogue in fade-in plugin.

Previously ply-entry was crashing with an assertion "entry->loop == NULL"
whenever the entry was updated. The change is basicly a copy of the working
version from spinfinity.

16 years agoexit plymouth-set-default-plugin cleanly when --rebuild-initrd isn't given
Ray Strode [Fri, 6 Mar 2009 23:03:28 +0000 (18:03 -0500)] 
exit plymouth-set-default-plugin cleanly when --rebuild-initrd isn't given

16 years agoDon't average the previous boot time with current boot time
Ray Strode [Fri, 6 Mar 2009 22:06:16 +0000 (17:06 -0500)] 
Don't average the previous boot time with current boot time

It's not always an improvement and it will always be worse
the second boot of a persistent live image.

16 years agoAlways allocate our own pty master when attaching
Ray Strode [Thu, 5 Mar 2009 19:39:37 +0000 (14:39 -0500)] 
Always allocate our own pty master when attaching

Previously, we'd let the caller pass one in, but
I don't think there's much point in allowing that.

16 years agoAllow -1 to get passed to ply-terminal-session
Ray Strode [Thu, 5 Mar 2009 19:38:28 +0000 (14:38 -0500)] 
Allow -1 to get passed to ply-terminal-session

This is coming from the drop-nash branch.  It lets
the caller of plymouth not have to create a pty master.

16 years agomake the glow plugin use its own icons
William Jon McCann [Wed, 4 Mar 2009 22:27:04 +0000 (17:27 -0500)] 
make the glow plugin use its own icons

Since the plugins may be packaged separately it isn't
kosher to share another plugins icons.

16 years agofix the help text for the color stops
William Jon McCann [Wed, 4 Mar 2009 22:10:57 +0000 (17:10 -0500)] 
fix the help text for the color stops

Make the help text use the proper argument name.

16 years agoadd a glow plugin
William Jon McCann [Wed, 4 Mar 2009 20:07:11 +0000 (15:07 -0500)] 
add a glow plugin

This plugin will progress through a sequence of images until
the progress reaches 90% complete.  At that point it will run
an animation sequence.

16 years agoadd a simple progress sequence helper
William Jon McCann [Wed, 4 Mar 2009 21:30:55 +0000 (16:30 -0500)] 
add a simple progress sequence helper

This is similar to the progress bar except that it uses a
sequence of images for the stages of the progress.

16 years agoadd simple one time animation helper
William Jon McCann [Wed, 4 Mar 2009 21:20:35 +0000 (16:20 -0500)] 
add simple one time animation helper

This will play a sequence of images from beginning to end at
a fixed frame rate.  It differs from the throbber in that it starts
at the 0th image and does not repeat.

16 years agouse the height in the max height comparison
William Jon McCann [Wed, 4 Mar 2009 03:28:57 +0000 (22:28 -0500)] 
use the height in the max height comparison

Fixes what was likely a copy/paste bug.

16 years agouse the configured colors for the gradient
William Jon McCann [Wed, 4 Mar 2009 02:54:53 +0000 (21:54 -0500)] 
use the configured colors for the gradient

Use the colors specified at configure time instead of
hard-coded values.

16 years agofix the trace messages when enabling/disabling logging
William Jon McCann [Thu, 26 Feb 2009 05:47:00 +0000 (00:47 -0500)] 
fix the trace messages when enabling/disabling logging

The sense of the messages was reversed.

16 years agoremove spurious disabling of console redirection
William Jon McCann [Thu, 26 Feb 2009 05:44:26 +0000 (00:44 -0500)] 
remove spurious disabling of console redirection

Revert accidental commit from:
6b9011d164201951af27f65163519b194d301572

16 years agoadd option to enable more compiler warnings
William Jon McCann [Tue, 24 Feb 2009 01:25:29 +0000 (20:25 -0500)] 
add option to enable more compiler warnings

Compiler warnings are helpful for finding latent bugs.

16 years agoadd way to detect if an option was set or not
William Jon McCann [Tue, 24 Feb 2009 01:09:25 +0000 (20:09 -0500)] 
add way to detect if an option was set or not

With integer options there's no obvious value to mean "unset" like there is for
string types (and to a lesser extent boolean types), so this commit
adds a mechanism.

16 years agoadd debug command line option
William Jon McCann [Tue, 24 Feb 2009 00:01:41 +0000 (19:01 -0500)] 
add debug command line option

Used to override the kernel command line and enabled debugging

16 years agoadd a message display method
William Jon McCann [Mon, 23 Feb 2009 23:04:03 +0000 (18:04 -0500)] 
add a message display method

Useful for showing messages like "Shutting down..." etc. A plugin
may choose not to support this feature.

16 years agoadd no-daemon command line option
William Jon McCann [Mon, 23 Feb 2009 19:39:27 +0000 (14:39 -0500)] 
add no-daemon command line option

Useful for debugging and running under gdb

16 years agoadd a mode commandline option
William Jon McCann [Mon, 23 Feb 2009 22:15:15 +0000 (17:15 -0500)] 
add a mode commandline option

Initially supports modes: boot, shutdown.  This allows the
progress cache to be loaded from the appropriate file.

16 years agoadd command line parsing to daemon
William Jon McCann [Mon, 23 Feb 2009 22:05:31 +0000 (17:05 -0500)] 
add command line parsing to daemon

If for no other reason than to handle --help.

16 years agoallow the progress cache file to be configurable
William Jon McCann [Sat, 21 Feb 2009 01:11:42 +0000 (20:11 -0500)] 
allow the progress cache file to be configurable

This will enable using separate cache files for different plymouth modes.

16 years agoimprove the layout of the help output
William Jon McCann [Fri, 20 Feb 2009 23:59:45 +0000 (18:59 -0500)] 
improve the layout of the help output

Left align descriptions and group subcommands.

16 years agoFix leak in ply-boot-server
Ray Strode [Tue, 24 Feb 2009 18:37:02 +0000 (13:37 -0500)] 
Fix leak in ply-boot-server

This commit fixes a memory leak
in an error path of ply_boot_connection_read_request.

16 years agoFix heap corruptor in ply-boot-server
Ray Strode [Tue, 24 Feb 2009 18:36:15 +0000 (13:36 -0500)] 
Fix heap corruptor in ply-boot-server

This fixes a memory corruption bug that
cropped up during a round of leak fixes
a while back.

16 years agoIn plymouth-set-default-plugin on reset make sure we do not set label.so as a
Charlie Brej [Tue, 24 Feb 2009 09:27:53 +0000 (09:27 +0000)] 
In plymouth-set-default-plugin on reset make sure we do not set label.so as a
default splash plugin.

16 years agoAdd hashtable and bitarray functionality
Charlie Brej [Sat, 3 Jan 2009 22:18:24 +0000 (22:18 +0000)] 
Add hashtable and bitarray functionality

Hashtable uses the bitarray to mark the used/dead places in the hash data array.
The hashtable functions follow the general gist of the glib ones. To create a
hashtable you supply the hashing and comparison functions. Examples of direct
and string based indexing are supplied. Hashtable is a single word because there
is no hash_somethingelse but the name may change soon. Notice the compare
function returns a signed int difference rather than an equal bool. This is
because it may be an idea to later migrate the collision system to a tree
sub-structure (currently it is a linear array search with an incrementing step
size).

16 years agoAllow client request answers to be greater than 255 characters.
Charlie Brej [Sat, 3 Jan 2009 21:33:56 +0000 (21:33 +0000)] 
Allow client request answers to be greater than 255 characters.

The new limit is 2^32. This only affects the replies to the client. Data sent
along with requests (e.g. password prompt) remain limited to <256 characters.

Two new functions have been added to the ply-utils (ply_read_uint32 and
ply_write_uint32) which receive/transmit 32bit numbers on FD streams.

AFAICT in function ply_boot_client_process_incoming_replies the answer was not
NUL terminated, which may have caused some of the password error bugs.

16 years agoOptionally pause progress while asking for a password or a question
Charlie Brej [Sat, 3 Jan 2009 19:41:21 +0000 (19:41 +0000)] 
Optionally pause progress while asking for a password or a question

The "Rework input methods" commit broke the pausing of the progress bar during
password entry. This commit reimplements that functionality but this time the
pausing is optional, applied with a --pause-progress parameter. The client
program sends the (un)pause commands before/after the password requests.
The default is currently to not pause but that may change quite soon.

This also fixes a couple minor bugs and formatting from "Rework input methods".
Also changed "progress-pause" to "pause-progress" to make it more human
intuitive.

16 years agoMake Ctrl+C and Ctrl+D cause the password/question entry to quit
Charlie Brej [Sat, 3 Jan 2009 15:58:39 +0000 (15:58 +0000)] 
Make Ctrl+C and Ctrl+D cause the password/question entry to quit

When asking for a password or a question, if the user presses Ctrl+C or Ctrl+D
the backend will reply a NULL. This is interprited as a "Stop asking me". The
client will not repeat the request (even if number of retries hasn't been
reached) and quit with an error exit_status.

16 years agoAdd progress-pause/unpause commands to halt the progress bar
Charlie Brej [Sun, 28 Dec 2008 18:34:25 +0000 (18:34 +0000)] 
Add progress-pause/unpause commands to halt the progress bar

Progress bar expansion was previously halted on password entry. Now the
progress can be halted whenever the system wishes. This can be useful when the
system notices there is a big task which is normally not present during normal
boots e.g. fsck.

16 years agoRework of the input methods
Charlie Brej [Fri, 19 Dec 2008 14:16:33 +0000 (14:16 +0000)] 
Rework of the input methods

Previously the splash plugins would deal with password entry, which would mean
that code was repeated and upon splash unload the password entry would be lost.
Now the keyboard strokes are processed with outside the splash, and the plugins
only deal with showing an appropriate user interface. As well as password
entry, the system can ask questions with non hidden text entry and sensing of
keystrokes.

16 years agoEnsure ply-buffer data block always terminates with a '\0'
Charlie Brej [Wed, 17 Dec 2008 12:07:57 +0000 (12:07 +0000)] 
Ensure ply-buffer data block always terminates with a '\0'

In some instances, ply-buffer did not terminate the data block with a '\0'.
Doing so allows the data to be used as a string. Additionally it now tries
harder to deal with very long additions to the buffer by repeatedly expanding
the capacity and dealing with appends larger than the maximum buffer size.

16 years agoPost release bump to 0.7.0
Ray Strode [Tue, 25 Nov 2008 19:21:04 +0000 (14:21 -0500)] 
Post release bump to 0.7.0

16 years agoAllow NULL to be passed to ply_image_free() 0.6.0
Ray Strode [Mon, 24 Nov 2008 22:42:11 +0000 (17:42 -0500)] 
Allow NULL to be passed to ply_image_free()

free() allows NULL so there is no reason ply_image_free
shouldn't.  This also papers over a bug in the solar
plugin that we haven't been able to identify yet.

16 years agoMerge branch 'master' of ssh://git.freedesktop.org/git/plymouth
Charles Brej [Fri, 21 Nov 2008 17:00:46 +0000 (17:00 +0000)] 
Merge branch 'master' of ssh://git.freedesktop.org/git/plymouth

16 years agoWhen fail when a command parser command is not in list then fail.
Charles Brej [Fri, 21 Nov 2008 16:58:25 +0000 (16:58 +0000)] 
When fail when a command parser command is not in list then fail.
Previously it returned the last in the list.

16 years agoCall --reset when giving default plugin if needed
Ray Strode [Fri, 21 Nov 2008 16:49:53 +0000 (11:49 -0500)] 
Call --reset when giving default plugin if needed

Previously, we tried to do this but failed because
of a buggy conditional.

16 years agoUpdate .gitignore
Ray Strode [Thu, 20 Nov 2008 02:31:08 +0000 (21:31 -0500)] 
Update .gitignore

There were a few bits of noise in git-status output

16 years agoword wrap README
Ray Strode [Wed, 19 Nov 2008 20:58:38 +0000 (15:58 -0500)] 
word wrap README

16 years agoUpdate README to reflect current state of things
Ray Strode [Wed, 19 Nov 2008 20:56:28 +0000 (15:56 -0500)] 
Update README to reflect current state of things

Some of the file paths and other information were out of date.

16 years agoAdd --list to plymouth-set-default-plugin
Ray Strode [Wed, 19 Nov 2008 19:18:48 +0000 (14:18 -0500)] 
Add --list to plymouth-set-default-plugin

https://bugs.freedesktop.org/show_bug.cgi?id=18298 requests
the ability to a get a list of plugins.

16 years agoAdd a --rebuild-initrd arg for set-default-plugin
Ray Strode [Wed, 19 Nov 2008 16:21:13 +0000 (11:21 -0500)] 
Add a --rebuild-initrd arg for set-default-plugin

Normally when a user runs plymouth-set-default-plugin
to change which plugin plymouth uses, the change doesn't
take effect until a new kernel is installed and the initrd
is rebuilt.

This new --rebuild-initrd argument forces the currently
running initrd to get rebuilt immediately (bug 18297).

16 years agoAdd usage info to plymouth-set-default-plugin
Ray Strode [Wed, 19 Nov 2008 15:40:40 +0000 (10:40 -0500)] 
Add usage info to plymouth-set-default-plugin

16 years agoTypo missing semicolon.
Charles Brej [Wed, 19 Nov 2008 16:22:01 +0000 (16:22 +0000)] 
Typo missing semicolon.

16 years agoUpdate progress as soon as the splash is shown
Charles Brej [Wed, 19 Nov 2008 16:18:08 +0000 (16:18 +0000)] 
Update progress as soon as the splash is shown

16 years agoDon't error if default.so doesn't exist
Ray Strode [Mon, 17 Nov 2008 20:31:59 +0000 (15:31 -0500)] 
Don't error if default.so doesn't exist

There are cases where we don't set a default and instead
fallback at runtime to text.so.  Don't make that an error
condition.

16 years agoOnly tell gdm to keep vt if told to retain splash
Ray Strode [Mon, 17 Nov 2008 15:47:08 +0000 (10:47 -0500)] 
Only tell gdm to keep vt if told to retain splash

See http://bugzilla.redhat.com/471785

16 years agoremove superfluous prototype
Ray Strode [Mon, 17 Nov 2008 14:54:41 +0000 (09:54 -0500)] 
remove superfluous prototype

16 years agoDon't loop forever when tty returns NUL bytes
Ray Strode [Fri, 14 Nov 2008 18:25:56 +0000 (13:25 -0500)] 
Don't loop forever when tty returns NUL bytes

Somehow a user was running into a case where plymouthd
would busy loop taking 100% cpu.  gdb revealed that it
was stuck trying to process keyboard input.  This is
apparently because we were looping forever when mbrlen()
got a NUL byte.

16 years agoTweak ply-progress to not progress too far beyond where it expects the next status...
Charles Brej [Mon, 17 Nov 2008 14:41:58 +0000 (14:41 +0000)] 
Tweak ply-progress to not progress too far beyond where it expects the next status update

This stops the progress from reaching 100% in cases such as fsck and timeouts.
It also averages the progress times from the previous one with the current one
to average out occasional slow tasks.

16 years agoGenerate background instead of using image
Charlie Brej [Fri, 14 Nov 2008 16:25:04 +0000 (11:25 -0500)] 
Generate background instead of using image

The background image previously was low resolution
and created various artifacts when being upscaled
to the native resolution of the panel.  This patch
ditches the background image and instead generates
a similar type of background dynamically, custom
fit to the screen it's being displayed on.  As an
added bonus, since we're doing it dynamically, we
can make some of the stars in the background
lightly twinkle.