OPIE applets: fix postinst() to not be used during image create
- usage of "pidof -s qpe" is useless on build machine and make dead processes here
- close bug #309
- creation of postinst()/postrm() should be moved to opie.bbclass
libqpe/libqpe-opie_1.2.1:
last minute backport from http://opie-bugs.oszine.de/view.php?id=1695 to
fix the font sizes for the Date selector on displays with higher
resolution.
packages/libqpe/libqpe-opie_1.2.1.bb:
The nomax patch was included in the main opie tree.
There is no need to apply it anymore.
packages/libqpe/libqpe-opie/fix-nomax.patch:
this patch is not needed anymore, this fix was applied
upstream.
Triton XScale Board Support:
Add latest patch courstey justin.huff BEI blueviewtech.com
from Bug-Report #316.
Do not hardcode the used FPU - leave that to the distribution
and starting using PV variable
conf/distro/openzaurus: rename openzaurus.conf to openzaurus.inc as it's not a valid DISTRO target. This causes builds to fail *before* anything is built.
Richard Purdie [Sun, 11 Sep 2005 22:10:49 +0000 (22:10 +0000)]
IMAGE_FSTYPE updates. Add new tar.bz2 and tar.gz image types so the compression format can be specified. Update tar references to tar.bz2. Update spitz and borzoi to correctly use tar.gz. ***Users of IMAGE_FSTYPE="tar" are advised to switch to tar.bz2***. tar is likely to become an uncompressed tarball at some point.
firefox: added the latest beta/rc 1.5b1
currently set at DEFAULT_PREFERENCE=-1 as it currently builds and runs but doesnt actually open a window on my OZ/GPE setup.
Richard Purdie [Sun, 11 Sep 2005 19:29:12 +0000 (19:29 +0000)]
zaurus-updater: The tar utility found on the Sharp inbuilt root filesystems used for reflashing the devices has bugs, one being that it won't create device files. We therefore need an external sane tar binary to successfully extract the root filesystem onto the device. This adds such a binary for spitz and borzoi.
conf/local.conf.sample: add a note on DISTRO and INHIBIT_PACKAGE_STRIP
local.conf.sample contained a wrong default which was masked by
http://bugs.openembedded.org/show_bug.cgi?id=319
John Bowler [Sat, 10 Sep 2005 18:22:30 +0000 (18:22 +0000)]
gcc_3.4.4: fix the native build on ucslugc by using arm, not thumb
For some reason a thumb built gcc 3.4.4 fails to produce functional
executables even when linked dynamically (i.e. even when next to none
of the code in the resultant executable comes from the compiler itself).
This requires further investigation (or maybe the csl gcc), this commit
changes the build to generate arm even on thumb systems as a temporary fix.
irda-utils:
-included /usr/include and /usr/src/linux/include via SYS_INCLUDES
variable. export a better candiate for SYS_INCLUDES
-version 0.9.15 is currently unbuildable, mark it as such
Opietooth:
do not install the headers with generic names like parser.h
in $STAGING_INCDIR but in a dedicated subdir
packages/libopietooth/libopietooth:
-create a common file (libopietooth1.inc)
-install the headers in a subdirectory
-adjust the 1.2.0 and cvs version to include the new file
packages/opie-bluetoothmanager/opie-bluetoothmanager:
-adjust to the new place for the header files
packages/opie-bluetoothapplet/opie-bluetoothapplet:
-adjust to the new place for the header files
Update Qt3 from Qt3.3.x to Qt3.3.5
packages/qt/qt-x11-free-native:
-update from 3.3.2 to 3.3.5
packages/qt/qt-x11-free:
-update from 3.3.3 to 3.3.5
packages/qte-fonts-common/qte-fonts-common:
-update from 3.3.3 to 3.3.5
packages/uicmoc/uicmoc3-native:
-update from 3.3.2 to 3.3.5
-build not the complete Qt3 but only most parts
-use ${PV} instead of hardcondig version
packages/qmake/qmake-native_1.07:
-update from 3.3.3 to 3.3.5
John Bowler [Fri, 9 Sep 2005 22:47:04 +0000 (22:47 +0000)]
uclibc, alsa, monotone-5, util-linux: enable thumb on ucslugc, disable it on these packages
This commit causes thumb build for everything in ucslugc-packages except
these packages and libgcc (from gcc). libgcc and uclibc still need some
work, the other packages reveal compiler bugs (monotone), source bugs
(alsa) or missing configuration (thumb arch support in util-linux) and so
are disabled for the moment.
John Bowler [Fri, 9 Sep 2005 22:38:20 +0000 (22:38 +0000)]
gcc(libgcc_s.so.1): fix the arm interwork _call_via_rX out of line ops
The interworking symbols were being built but not exported from libgcc
(i.e. they were not in the version file), added a specific version file
for them and updated t-linux. Forced the cross build to always generate
ARM instructions (the thumb version does not currently work), bumped
PRs.
Richard Purdie [Fri, 9 Sep 2005 21:46:17 +0000 (21:46 +0000)]
linux-oz-2.6: Rehash spitz patches after advice from mainline. Add delay to mmc_detect_change() and update corgi and spitz to use this. hx2750 and tosa need updating. More patches pushed. hrw moved the collie/poodle keymap into initscripts so the collie.keymap patch is no longer needed.
John Bowler [Fri, 9 Sep 2005 18:23:50 +0000 (18:23 +0000)]
mt-daapd: fix 0.2.1.1 to work with 'iTunes5'
The fix applies a patch copied from www.mt-daapd.org, the patch is copied
because the version at the web site has path names (well, no path names)
which don't work with OE. This commit also adds a .bb for mt-daapd_0.2.2,
but since that hasn't been released it doesn't work (yet) and is DPed out.
John Bowler [Fri, 9 Sep 2005 01:47:04 +0000 (01:47 +0000)]
uclibc_svn: fix the SVN version to build the uclibc SVN head
The SVN version now copies 0.9.28 patches, the armeb case should work,
other architectures almost certainly require the uclibc-cvs/*/uClibc.config
copy of the config file to be deleted (if this is done the build will pick
up the config file in uclibc-0.9.28).
John Bowler [Fri, 9 Sep 2005 01:39:02 +0000 (01:39 +0000)]
uclibc: fix STT_ARM_TFUNC handling in the uclibc ld.so
STT_ARM_TFUNC handling was broken in uclibc because it is an ARM specific
hack and uclibc code did not have support for it. Added the support in a
way which I hope will be acceptable to uclibc and which should remain
compatible with forthcoming binutils changes. ARM Thumb code is fully
working with this commit with the possible exception of uclibc/gcc itself
(not yet tested).