Age | Commit message (Collapse) | Author | Files |
|
functionality.
* Essentially, a bootloader with rootfs location selection. Only rootfs for already
booted kernel, no kexec support.
* Note: *generic* boot menu/loader, not some "zauri" mumbo-jumbo.
|
|
boot.
* Note: *allows*, not starts. To do that, "dropbear=early" should be passed on
kernel command line.
|
|
* Currently in Angstrom we have to "Keyboard" IMs. Not good.
* Corrections for better title welcome.
|
|
* These files weren't used before, and now are just stale, after logo
image cleanup.
|
|
* So, psplash uses bottom 1/6 of screen for progress bar. So, teach
it to use 5/6 of the screen height for logo placement, not the whole
height, which leads to overlap of logo and bar.
|
|
* This one now looks perfectly for QVGA and higher resolutions - proper size,
proper background.
|
|
* Add timer to shutdown psplash after 120 seconds. Normal boot procedure should
be over by then, but if not, user at least will see on the console screen where
it hung.
|
|
* This one directly suitable for use as psplash logo, having optimal
size for QVGA resolution, and of course working well for any higher
resolution too (will be centered on the screen).
|
|
* Anyway, what should be used is design marked as "Small version for psplash".
For use as psplash background, it should be exported with width 230 pixels
(sic, this is teh width which fits well with the progress indicator).
|
|
|
|
* Also, designate start and end of "ipkg configure" run, to diagnose
boot-time hangs.
|
|
* Don't spit detailed info about links created unless passed -v.
* This is to prevent initial bootup console spamming.
|
|
|
|
* Icon was shamelessly hacked up from matchbox-keyboards' and gpe-conf-time's.
|
|
official.
* Contributed by Jeroen Hoek <mail at jeroenhoek.nl>, see
http://lists.linuxtogo.org/pipermail/angstrom-distro-devel/2007-May/000286.html
|
|
and 'cefef520b1840df3899cd9fc5a4fc3cb9faa38ce'
|
|
Suggested by Richard as a way to force ipk's of dependencies
to get build when building a package. This is handy if
your are building a package for a feed that does not get
installed in an image.
|
|
and 'f7531c375389681cc06ea3169fea009301b94267'
|
|
|
|
busybox ifup doesn't support allow-hotplug.
|
|
and '0dbb48946ab15c0ee9c232d1a1fc31e5b65037cd'
|
|
Angstrom (or other DISTRO) flashable images without needing to create a new DISTRO
|
|
the later stages like install/packaging/deploy can make use of this. This
was done so that kernel-image could contain a uImage but is suitable for
other cases as well.
|
|
Move eglibc to revision 4355 which has -Os support.
|
|
|
|
and '9c32fcb38cb012f79311cd1b07560964873c6307'
|
|
|
|
and '332e97250454514e2c07aea9f1695ab21166d191'
|
|
and 'a772b6bc0f51e8c371cc638bb29f215e00e02bc5'
|
|
Landscape / Portrait -> Portrait rotataions)
- Power off display on lid-close and remember old brightness setting
- Run power-on function in the background to allow faster response on lid-open
|
|
|
|
and '805d2f01c6463550c4d0dd1f943fc6f32ae13613'
|
|
and 'dd83c0ba4e83a6207462418b614431ce061d9481'
|
|
|
|
|
|
|
|
|
|
* When autolock was activated in lanscape mode, and then switched to
portrait, the window doesn't fit to screen. Fixed.
|
|
the mtd partitions from usespace so disable for now.
|
|
* One of the very noticeable pain points in matchbox' IM handling is high
latency of IM popping up, which can take up to half-second with complex
screen layouts. This is rooted in the fact that IM process is killed when
IM window is hidden, and started again afterwards.
* I wanted to address this on inputmanager level, transparently to actual
IM implementation, by hiding/showing IM's window by the manager. Unfortunately,
there appears to be no reliable method to figure out window IDs from process
properties (for example, _MB_CLIENT_EXEC_MAP X property, maintained by matcbox,
doesm't capture all apps started, in particular, it captures neither xkbd
nor matchbox-keyboard, GPE's term app is not caught too, while either GPE apps
are). Then, wanted to implement custom PropertyNotify-based protocol, though
there would be left issue of distinguishing IMs which support this protocol
and ones which don't. Finally, by random gerpping, I found that xkbd supports
Unix signal based visibility protocol! By receiving SIGUSR1, it hides/shows
its window. This seems like 100% win, as the default behavior for SIGUSR1
is process termination, so unsupporting clients will be just terminated as
before.
|
|
|
|
|
|
|
|
|
|
|
|
2.6.20.20
|
|
|
|
closes #3426
|
|
and '33e5bd40976428aea7682da1e3cb6ed9c93af23e'
|
|
unused alternatives for cmdline - these recipes are for nslu2 only.
|