scm/bb/tizen-distro.git
13 years agoimagetest-qemu: Add test case for dmesg check in target
Jiajun Xu [Fri, 29 Oct 2010 17:03:22 +0000 (01:03 +0800)]
imagetest-qemu: Add test case for dmesg check in target

Add a test case for error log check with command dmesg in target. The
case introduces a new folder in target, "/opt/test", which holds test
scripts running in target.

Signed-off-by Jiajun Xu <jiajun.xu@intel.com>

13 years agopoky.conf: Change default locale to POSIX
Saul Wold [Thu, 4 Nov 2010 18:01:42 +0000 (11:01 -0700)]
poky.conf: Change default locale to POSIX

Signed-off-by: Saul Wold <Saul.Wold@intel.com>
13 years agotask-sdk-gmae: fix typo in SDK-EXTRAS and convert to ?= for assignment
Saul Wold [Thu, 4 Nov 2010 17:49:29 +0000 (10:49 -0700)]
task-sdk-gmae: fix typo in SDK-EXTRAS and convert to ?= for assignment

Signed-off-by: Saul Wold <Saul.Wold@intel.com>
13 years agopython: Fix hardcoded python references which should refer to PN, removing duplicate...
Richard Purdie [Wed, 10 Nov 2010 13:07:12 +0000 (21:07 +0800)]
python: Fix hardcoded python references which should refer to PN, removing duplicate providers warning

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agobusybox: import upstream patch for make 3.82
Joshua Lock [Tue, 9 Nov 2010 12:18:14 +0000 (12:18 +0000)]
busybox: import upstream patch for make 3.82

Signed-off-by: Joshua Lock <josh@linux.intel.com>
13 years agoeglibc: fix build of eglibc-initial for make 3.82
Joshua Lock [Tue, 9 Nov 2010 12:14:28 +0000 (12:14 +0000)]
eglibc: fix build of eglibc-initial for make 3.82

Make 3.82, as shipped with Fedora 14, fixes some holes in the parser which in
turn breaks behaviour of some Makefiles. Most notably eglibc's.

Signed-off-by: Joshua Lock <josh@linux.intel.com>
13 years agopackage.bbclass: make sure 'sysroots' created before lockfile
Kevin Tian [Tue, 26 Oct 2010 07:54:43 +0000 (15:54 +0800)]
package.bbclass: make sure 'sysroots' created before lockfile

package sstate requires a lock under sysroots/, which however may not be
created when sstate_setscene functions are executed and then causes failures.
here we make sure 'sysroots' created before do_package_setscene is executed.

Signed-off-by: Kevin Tian <kevin.tian@intel.com>
13 years agopoky.conf: add PWD to BB_HASHBASE_WHITELIST
Kevin Tian [Fri, 5 Nov 2010 07:16:50 +0000 (15:16 +0800)]
poky.conf: add PWD to BB_HASHBASE_WHITELIST

Without doing this I can't reuse prebuilts generated from one directory in another
build directory which is really not comfortable for developers. Due to difference
on PWD variable, base hash becomes different for every sstate package.

So here PWD is added to the hase base whitelist, which makes most prebuilts reusable
now in a different build directory. However there still remains one issue about
do_populate_sysroot sstate package. Absolute path is encoded in this task which
then blocks reuse on them. Perhaps we need some redesign for do_populate_sysroot into
two stages, with all absolute path related operations moved to the 2nd stage. This
way then the 1st stage can be reused.

Partially fix [BUGID #522]

Signed-off-by Kevin Tian <kevin.tian@intel.com>

13 years agobitbake.runqueue.py: Fix no exec task handling to correct counter housekeeping
Richard Purdie [Sun, 7 Nov 2010 20:33:12 +0000 (20:33 +0000)]
bitbake.runqueue.py: Fix no exec task handling to correct counter housekeeping

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agometa/classes: Mark do_package_write and do_build as noexec tasks
Richard Purdie [Sat, 6 Nov 2010 14:33:45 +0000 (14:33 +0000)]
meta/classes: Mark do_package_write and do_build as noexec tasks

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agobitbake: Add support for 'noexec' tasks
Richard Purdie [Sat, 6 Nov 2010 14:33:05 +0000 (14:33 +0000)]
bitbake: Add support for 'noexec' tasks

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agobitbake/siggen: Allow siggen code to control final stampfile format
Richard Purdie [Sat, 6 Nov 2010 12:20:33 +0000 (12:20 +0000)]
bitbake/siggen: Allow siggen code to control final stampfile format

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years ago.gitignore: Update after documentation layout changes
Richard Purdie [Sat, 6 Nov 2010 12:19:51 +0000 (12:19 +0000)]
.gitignore: Update after documentation layout changes

13 years agocross-canadian bbclass: replace hardcoded -pokysdk with SDK_VENDOR
Koen Kooi [Wed, 3 Nov 2010 08:48:43 +0000 (09:48 +0100)]
cross-canadian bbclass: replace hardcoded -pokysdk with SDK_VENDOR

Signed-off-by: Koen Kooi <k-kooi@ti.com>
Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agoopenssl: Use STAGING_BINDIR_NATIVE as the path to perl, not bindir which can point...
Richard Purdie [Wed, 3 Nov 2010 10:06:02 +0000 (10:06 +0000)]
openssl: Use STAGING_BINDIR_NATIVE as the path to perl, not bindir which can point to incorrect places

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agosstate: Provide a temporary mechanism to stop collisions between toolchain components...
Richard Purdie [Tue, 2 Nov 2010 13:18:07 +0000 (13:18 +0000)]
sstate: Provide a temporary mechanism to stop collisions between toolchain components in builds from scratch using sstate

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agobitbake/siggen: Provide better warnings about empty tasks
Richard Purdie [Tue, 2 Nov 2010 13:17:06 +0000 (13:17 +0000)]
bitbake/siggen: Provide better warnings about empty tasks

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agoexternal-python-tarball: Include SDKMACHINE in tarball name
Richard Purdie [Tue, 2 Nov 2010 13:15:42 +0000 (13:15 +0000)]
external-python-tarball: Include SDKMACHINE in tarball name

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agoexternal-python-tarball: Add ability to generate an externally usable standalone...
Richard Purdie [Tue, 2 Nov 2010 13:11:37 +0000 (13:11 +0000)]
external-python-tarball: Add ability to generate an externally usable standalone python tarball

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agoopenssl/db/gdbm: Extend to provide BBCLASSEXTEND nativesdk
Richard Purdie [Tue, 2 Nov 2010 13:10:36 +0000 (13:10 +0000)]
openssl/db/gdbm: Extend to provide BBCLASSEXTEND nativesdk

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agopython: Remove hardcoded python references and replace with PN, allowing use of BBCLA...
Richard Purdie [Tue, 2 Nov 2010 13:08:49 +0000 (13:08 +0000)]
python: Remove hardcoded python references and replace with PN, allowing use of BBCLASSEXTEND=nativesdk

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agobase/utils.bbclass: Sync up overlapping functions
Richard Purdie [Tue, 2 Nov 2010 11:50:50 +0000 (11:50 +0000)]
base/utils.bbclass: Sync up overlapping functions

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agoutils.bbclass: Add functions from OE.dev
Richard Purdie [Tue, 2 Nov 2010 10:33:52 +0000 (10:33 +0000)]
utils.bbclass: Add functions from OE.dev

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agognome.bbclass: Remove autotools_stage reference
Richard Purdie [Tue, 2 Nov 2010 10:15:27 +0000 (10:15 +0000)]
gnome.bbclass: Remove autotools_stage reference

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agognome.bbclass: Use BPN instead of PN and set SRC_URI name similar to OE.dev
Richard Purdie [Tue, 2 Nov 2010 10:11:05 +0000 (10:11 +0000)]
gnome.bbclass: Use BPN instead of PN and set SRC_URI name similar to OE.dev

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agometa-emenlow: Rename packages to recipes in line with main repository
Richard Purdie [Tue, 2 Nov 2010 10:08:00 +0000 (10:08 +0000)]
meta-emenlow: Rename packages to recipes in line with main repository

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agobase-files: Allow a basic issue file to be extended by the code, following the method...
Richard Purdie [Tue, 2 Nov 2010 09:20:46 +0000 (09:20 +0000)]
base-files: Allow a basic issue file to be extended by the code, following the method now used by OE

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agobase-files: Allow issue and hostname generation to be overridden
Richard Purdie [Tue, 2 Nov 2010 09:08:11 +0000 (09:08 +0000)]
base-files: Allow issue and hostname generation to be overridden

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agosrc_distribute: Set EXPORT_FUNCTIONS correctly
Richard Purdie [Mon, 1 Nov 2010 10:48:43 +0000 (10:48 +0000)]
src_distribute: Set EXPORT_FUNCTIONS correctly

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agobitbake/siggen: Fix finalise call arguments
Richard Purdie [Mon, 1 Nov 2010 10:23:09 +0000 (10:23 +0000)]
bitbake/siggen: Fix finalise call arguments

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agoPoky -> Yocto name change
Richard Purdie [Wed, 27 Oct 2010 06:54:03 +0000 (07:54 +0100)]
Poky -> Yocto name change

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agorm_work.bbclass: Handle case where pseudo directory doesn't exist
Richard Purdie [Fri, 22 Oct 2010 23:56:38 +0000 (00:56 +0100)]
rm_work.bbclass: Handle case where pseudo directory doesn't exist

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agobitbake/fetch: When fetchers return errors, ensure any partial download is cleared
Richard Purdie [Fri, 22 Oct 2010 23:22:54 +0000 (00:22 +0100)]
bitbake/fetch: When fetchers return errors, ensure any partial download is cleared

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agopackage_deb: The packaging command itself is run under fakeroot so these lines are...
Richard Purdie [Fri, 22 Oct 2010 22:05:45 +0000 (23:05 +0100)]
package_deb: The packaging command itself is run under fakeroot so these lines are totally unneeded

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agoFix adding new contact failure when starting contacts for the 1st time
Zhai Edwin [Fri, 22 Oct 2010 09:11:26 +0000 (17:11 +0800)]
Fix adding new contact failure when starting contacts for the 1st time

When contacts called for the 1st time, adding new contact always failed but
worked after restart. The reason is e-addressbook-factory used to handle
addressbook leads a segmentation fault. Root cause is that factory would
create the data base if not exist, but use it without proper
initialization.

[BUGID #455] got fixed by this.

Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
13 years agoclasses: Only enable fakeroot on setscene tasks with packaging
Richard Purdie [Fri, 22 Oct 2010 15:34:06 +0000 (16:34 +0100)]
classes: Only enable fakeroot on setscene tasks with packaging

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agobitbake/fetch: Make URL checking slightly less verbose (distracting with the sstate...
Richard Purdie [Fri, 22 Oct 2010 13:31:18 +0000 (14:31 +0100)]
bitbake/fetch: Make URL checking slightly less verbose (distracting with the sstate code)

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agosstate: Fix mirror handling for file:// urls
Richard Purdie [Fri, 22 Oct 2010 13:27:44 +0000 (14:27 +0100)]
sstate: Fix mirror handling for file:// urls

The fetcher has special handling for file:// mirror urls, being efficient and
just providing an updated path. Unfortunately the sstate fetching code wasn't
able to handle this. This patch detects this and injects a symlink to ensure
everything works. It also fixes some datastore references to be correct and
ensures the sstate download directory exists if it doesn't already.

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agopseudo/fakeroot: Move the pseudo directory creation into bitbake
Richard Purdie [Fri, 22 Oct 2010 12:00:40 +0000 (13:00 +0100)]
pseudo/fakeroot: Move the pseudo directory creation into bitbake

If sstate was used to accelerate a build, the pseudo directory might not have
been created leading to subsequent task failures.

Also, sstate packages were not being installed under pseudo context meaning
file permissions could have been lost.

Fix these problems by creating a FAKEROOTDIRS variable which bitbake ensures
exists before running tasks and running the appropriate setscene tasks under
fakeroot context.

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agopackage_deb: Fix a typo meaning the debian packaging was not running in the fakeroot...
Richard Purdie [Fri, 22 Oct 2010 11:59:54 +0000 (12:59 +0100)]
package_deb: Fix a typo meaning the debian packaging was not running in the fakeroot evnironment

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agopackage_rpm: Don't check for the existence of dvar as its never used
Richard Purdie [Fri, 22 Oct 2010 11:56:20 +0000 (12:56 +0100)]
package_rpm: Don't check for the existence of dvar as its never used

If a sstate package exists for the package task but not for the rpm packaging
task, the output from the package task will be used. The directory pointed
to by dvar will not exist under this scenario.

Since the directory is never used by the packaging process remove the
check, substituting the pkgd variable which is always present and used.

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agobase.bbclass: Ensure an empty do_build tasks exists to silence a warning
Richard Purdie [Fri, 22 Oct 2010 11:53:36 +0000 (12:53 +0100)]
base.bbclass: Ensure an empty do_build tasks exists to silence a warning

The message "WARNING: Function do_build doesn't exist" doesn't look professional,
so fix the underlying problem even if this warning is harmless.

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agosstate: Fix broken plaindirs support
Richard Purdie [Fri, 22 Oct 2010 11:51:35 +0000 (12:51 +0100)]
sstate: Fix broken plaindirs support

When installing a sstate package, directories tracked by plaindirs were being installed
to the incorrect location. With the current implementation this was limited to
the do_package task.

This patch ensures plaindirs tracked files are created in the correct location, fixing
the bug where these files would go missing.

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agoat: Fix nonstripbinaries.patch
Mark Hatle [Thu, 21 Oct 2010 17:26:14 +0000 (12:26 -0500)]
at: Fix nonstripbinaries.patch

The commit '0e0d14ed1e88ebff876abb3d00b08627a3879e8d' introduced a new patch
nonstripbinaries.patch, the filename format caused the patch to not apply.
Simple fix made by correcting the path.

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
13 years agolibtheora: add DEPENDS on libogg
Dexuan Cui [Thu, 21 Oct 2010 06:54:34 +0000 (14:54 +0800)]
libtheora: add DEPENDS on libogg

This is used to fix the following build failure:

 checking for oggpackB_read... no
| configure: error: newer libogg version (1.1 or later) required

Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
13 years agobase/sstate: Add cleanall task to remove downloads and sstate cached files
Richard Purdie [Fri, 22 Oct 2010 09:22:34 +0000 (10:22 +0100)]
base/sstate: Add cleanall task to remove downloads and sstate cached files

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agolocal.conf.sample: Default to not building 32 bit libs on 64 bit systems as most...
Richard Purdie [Thu, 21 Oct 2010 22:28:14 +0000 (23:28 +0100)]
local.conf.sample: Default to not building 32 bit libs on 64 bit systems as most people don't need it and it confuses them

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agometadata_scm: Ensure that if an SCM isn't present, we dont print a revision of 'fatal...
Richard Purdie [Thu, 21 Oct 2010 22:26:40 +0000 (23:26 +0100)]
metadata_scm: Ensure that if an SCM isn't present, we dont print a revision of 'fatal:' as it looks bad

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agobusybox/sysklogd: syslog.conf files
Mark Hatle [Wed, 20 Oct 2010 19:50:50 +0000 (14:50 -0500)]
busybox/sysklogd: syslog.conf files

The format of the syslog.conf files is different between busybox and sysklogd.

Use the alternatives method to ensure we get the correct config file for
any specific configurations.

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
13 years agogst-ffmpeg: powerpc build errors
Bruce Ashfield [Thu, 21 Oct 2010 04:37:54 +0000 (00:37 -0400)]
gst-ffmpeg: powerpc build errors

Fixes [BUGID #501]

While building the poky-image-sato target, gst-ffmpeg failed when
it attempted to run altivec tests on the host. We can live without
this optimization, so the easiest fix is to simply disable it.

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
13 years agompc8315e-rdb: align PACKAGE_EXTRA_ARCHS with tuning
Bruce Ashfield [Thu, 21 Oct 2010 04:30:07 +0000 (00:30 -0400)]
mpc8315e-rdb: align PACKAGE_EXTRA_ARCHS with tuning

Fixes [BUGID #500]

While the tuning for the mpc8315e is 603e, the PACKAGE_EXTRA_ARCHES
was ppce300. This created a mismatch and resulted in rootfs assembly
issues due to missing locales.

We align both at 603 and can revist a better tuning in the future.

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
13 years agoat: remove nonrootinstall patch and modify some relevant files
Mei Lei [Thu, 21 Oct 2010 04:36:51 +0000 (12:36 +0800)]
at: remove nonrootinstall patch and modify some relevant files

remove nonrootinstall patch, and use a patch named nonstripbinaries to modi
fy the /at-3.1.10.2/Makefile.in, so it can also handles stripping of the bi
naries

Fix [BUG 246]

Signed-off-by: Mei Lei <lei.mei@intel.com>
13 years agognome-icon-theme: fix build
Joshua Lock [Wed, 20 Oct 2010 21:25:22 +0000 (22:25 +0100)]
gnome-icon-theme: fix build

Duplicate the patch from sato-icon-theme which enables the path to the
icon-name-mapping tool to be specified.

Signed-off-by: Joshua Lock <josh@linux.intel.com>
13 years agobash: There is no such configure option as --with-ncurses and --with-curses breaks...
Richard Purdie [Wed, 20 Oct 2010 20:55:02 +0000 (21:55 +0100)]
bash: There is no such configure option as --with-ncurses and --with-curses breaks the build

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agopoky-default-revisions: define SRCREV for pseudo-nativesdk
Joshua Lock [Tue, 19 Oct 2010 12:22:56 +0000 (13:22 +0100)]
poky-default-revisions: define SRCREV for pseudo-nativesdk

Signed-off-by: Joshua Lock <josh@linux.intel.com>
13 years agonfs-utils: fix nfs server initscript
Qing He [Wed, 20 Oct 2010 08:49:04 +0000 (16:49 +0800)]
nfs-utils: fix nfs server initscript

fixes [BUGID #483]

Signed-off-by: Qing He <qing.he@intel.com>
13 years agobash: fix job control
Qing He [Wed, 20 Oct 2010 08:45:23 +0000 (16:45 +0800)]
bash: fix job control

job control can't be automatically detected when cross-compiling,
enable it explicitly.

Fixes [BUGID #487]

Signed-off-by: Qing He <qing.he@intel.com>
13 years agotask-sdk: Fix lttng-ust dependency to include the -dev version so headers are included
Richard Purdie [Tue, 19 Oct 2010 19:52:49 +0000 (20:52 +0100)]
task-sdk: Fix lttng-ust dependency to include the -dev version so headers are included

[BUGID #425]

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agosysklogd: correct the syslog link and conf file
Yu Ke [Tue, 19 Oct 2010 07:15:17 +0000 (15:15 +0800)]
sysklogd: correct the syslog link and conf file

This commit fix [BUGID #482]

Two issues cause bug 482:
- firsty, there are two version of syslog: sysklogd and busybox.
  the busybox one is directly installed as /etc/init.d/syslog,
  and the sysklogd one is installed by update-alternative. the
  update-alternative will thus fail because the /etc/init.d/syslog
  (busybox one) already exist and not a link. so the correct way
  is to install busybox one by update-alternative, the layout will be:
  /etc/init.d/syslog.busybox
  /etc/init.d/syslog.sysklogd
  /etc/init.d/syslog -> syslog.busybox or
  /etc/init.d/syslog -> syslog.sysklogd

- secondly, sysklogd default conf is not comply with poky. Its dir /var/adm/
  does not exist. Check the debian /etc/syslog.conf and find it is more
  sophiscated and suitable, so port /etc/syslog.conf from debian.

Signed-off-by: Yu Ke <ke.yu@intel.com>
13 years agowrs_meta: emenlow updates
Bruce Ashfield [Mon, 18 Oct 2010 19:33:30 +0000 (15:33 -0400)]
wrs_meta: emenlow updates

Fixes [BUGID #465]

    wrs_meta: add logbuf size feature

    A 16k kernel log buffer size is too small for emenlow and atom-pc -
    boot messages are routinely lost with the current setting.  Increase
    the size to something more reasonable - 64k seems to be a good
    compromise.

Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>
    emenlow: add cmdline option for reboot quirk

    The default reboot options don't work for emenlow i.e. kbd + triple
    results in a reboot hang.  According to the emenlow chipset manual,
    port CF9 is supported, so let's use that.

Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
13 years agobitbake/fetch/git: Ensure fullclone repositories are fully fetched
Richard Purdie [Tue, 19 Oct 2010 16:17:41 +0000 (17:17 +0100)]
bitbake/fetch/git: Ensure fullclone repositories are fully fetched

The git fetcher was failing to pull in new branches into a git
repository mirror tarball as the git fetch command being used didn't
add new remote branches.

This patch uses "git fetch --all" for fullclones to ensure any
new remote branches are cloned correctly.

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agobitbake/fetcher: Deal with a ton of different bugs
Richard Purdie [Tue, 19 Oct 2010 11:30:21 +0000 (12:30 +0100)]
bitbake/fetcher: Deal with a ton of different bugs

The more we try and patch up the fetcher code, the more things break. The
code blocks in question are practically unreadable and are full of corner
cases where fetching could fail. In summary the issues noticed included:

a) Always fetching strange broken urls from the premirror for "noclone"
   git repositories
b) Not creating or rewriting .md5 stamp files inconsistently
c) Always fetching git source mirror tarballs from the premirror even
   if they already exist but the checkout directory does now
d) Passing "None" values to os.access() and os.path.extsts() checks under
   certain circumstances
e) Not using fetched git mirror tarballs if the preexist and always
   try and fetch them.

This patch rewrites the sections of code in question to be simpler and
more readable, fixing the above problems and most likely other odd
corner cases.

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agobitbake/fetch/git.py: Fix git fetcher to correctly use mirror tarballs
Richard Purdie [Mon, 18 Oct 2010 23:12:09 +0000 (00:12 +0100)]
bitbake/fetch/git.py: Fix git fetcher to correctly use mirror tarballs

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agorpm: Change pcre to use internal library [BUGID #404]
Saul Wold [Sun, 17 Oct 2010 18:08:05 +0000 (11:08 -0700)]
rpm: Change pcre to use internal library [BUGID #404]

Signed-off-by: Saul Wold <Saul.Wold@intel.com>
13 years agosed: use ALTERNATIVE_LINK
Kevin Tian [Mon, 18 Oct 2010 09:38:32 +0000 (17:38 +0800)]
sed: use ALTERNATIVE_LINK

update-alternatives.bbclass assumes /usr/bin/sed if ALTERNATIVE_LINK is not
specified, however sed has changed its default installation directory to
/bin/sed. This causes below errors at the first boot of lsb image:

update-alternatives: Error: cannot register alternative sed to /usr/bin/sed
since it is already registered to /bin/sed

This fixes [BUGID #478]

Signed-off-by: Kevin Tian <kevin.tian@intel.com>
13 years agogstreamer: add support for theora plugins
Dongxiao Xu [Mon, 18 Oct 2010 02:51:56 +0000 (10:51 +0800)]
gstreamer: add support for theora plugins

Add theora codec and its corresponding lib recipe (import from OE).
This fixes [BUGID #476].

Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
13 years agowrs_meta: update SRCREV for netfilter options
Bruce Ashfield [Mon, 18 Oct 2010 05:15:33 +0000 (01:15 -0400)]
wrs_meta: update SRCREV for netfilter options

Fixes [BUGID #199]

Updating the SRCREV for wrs_meta to import a set of default
netfilter options, which can then be applied to any BSP. This
keeps the default BSP configs small, while allowing the
functionality to be available.

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
13 years agokernel tools: update SRCREV
Bruce Ashfield [Mon, 18 Oct 2010 05:12:08 +0000 (01:12 -0400)]
kernel tools: update SRCREV

Fixes [BUGID #199]

To allow optional features to be enabled for a given BSP, we
need to re-enable this functionality in the kernel tools. The
following commit in that repo fixes a problem when passing
feature names:

  [
  commit 140693a6b0d81b7ba7175b6cfce11c6c22f81e24
  Author: Bruce Ashfield <bruce.ashfield@windriver.com>
  Date:   Sun Oct 17 01:22:00 2010 -0400

    updateme: do not expand directories

    If a directory is passed to updateme, do NOT expand it to a
    full path, the compilation phase will search and do this
    while applying some special logic to find properly named
    features.

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
  ]

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
13 years agolinux-wrs: pass netfilter options as an addon feature
Bruce Ashfield [Sun, 17 Oct 2010 05:18:48 +0000 (01:18 -0400)]
linux-wrs: pass netfilter options as an addon feature

Fixes [BUGID #199]

Although an optional feature to a base BSP, most of the machines
currently supported want/need the netfilter options. To enforce
this configuration, it is temporarily forced in the kernel
recipe itself, but can move to machine configurations in the
future.

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
13 years agolinux-wrs: allow optional features via KERNEL_FEATURES variable
Bruce Ashfield [Sun, 17 Oct 2010 05:14:11 +0000 (01:14 -0400)]
linux-wrs: allow optional features via KERNEL_FEATURES variable

Fixes [BUGID #199]

Reinstate the ability to specify optional/additional kernel features
when updating the tree.

This is done via the variable KERNEL_FEATURES which specifies
a list of features to be appended to the current branch and
config. These features are part of the wrs_meta branch in the
kernel repository and hence are self contained within the
kernel tree waiting to be activated. This saves multiple
branches simply to allow a machine to have many profiles.

The kernel patching/configuration phases will locate these
features and add them to the meta_series, which in turn
modifies the tree.

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
13 years agodistrodata: Modify Distro tracking spreadsheet
Saul Wold [Fri, 15 Oct 2010 20:51:33 +0000 (13:51 -0700)]
distrodata: Modify Distro tracking spreadsheet

Signed-off-by: Saul Wold <Saul.Wold@intel.com>
13 years agodistro_tracking: Update Distro Aliases
Saul Wold [Fri, 15 Oct 2010 20:50:34 +0000 (13:50 -0700)]
distro_tracking: Update Distro Aliases

Signed-off-by: Saul Wold <Saul.Wold@intel.com>
13 years agotask-poky-sdk: Add lttng-ust
Richard Purdie [Thu, 14 Oct 2010 19:20:23 +0000 (20:20 +0100)]
task-poky-sdk: Add lttng-ust

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agolibtasn1: Comment out the RREPLACES as it causes problems with rpm backend
Richard Purdie [Fri, 15 Oct 2010 15:38:24 +0000 (16:38 +0100)]
libtasn1: Comment out the RREPLACES as it causes problems with rpm backend

Since we've never distributed binary feeds this isn't a problem for now
for Poky and as it stands this line breaks images.

[BUGID #442]

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agopoky.conf: Add qemu-config dependency for qemux86-64
Richard Purdie [Fri, 15 Oct 2010 15:34:40 +0000 (16:34 +0100)]
poky.conf: Add qemu-config dependency for qemux86-64

After booting qemux86-64 sato image, there is no icon for shutdown on the
desktop which is caused by missing the qemu-config package on this machine.

[BUGID #458]

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agoshared-mime-info: Don't run the util for the native version of the package
Richard Purdie [Fri, 15 Oct 2010 15:30:41 +0000 (16:30 +0100)]
shared-mime-info: Don't run the util for the native version of the package

The whole point of building a native version of this package is
to provide the share-mime-info binary so using it in do_install
of the native package is incorrect. Add a modified do_install for
the native version of the package to avoid this.

[BUGID #466]

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agolibffi: fix the uri
Joshua Lock [Fri, 15 Oct 2010 13:42:34 +0000 (14:42 +0100)]
libffi: fix the uri

It had the previous version hard-coded

Signed-off-by: Joshua Lock <josh@linux.intel.com>
13 years agodocumentation: Complete poky-handbook -> poky-ref-manual rename
Richard Purdie [Fri, 15 Oct 2010 11:19:34 +0000 (12:19 +0100)]
documentation: Complete poky-handbook -> poky-ref-manual rename

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agohandbook: Move into documentation directory
Richard Purdie [Fri, 15 Oct 2010 10:55:59 +0000 (11:55 +0100)]
handbook: Move into documentation directory

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agoUse gdk_threads_enter/gdk_threads_leave to ensure gtk multithread safe
Dongxiao Xu [Fri, 15 Oct 2010 08:45:21 +0000 (16:45 +0800)]
Use gdk_threads_enter/gdk_threads_leave to ensure gtk multithread safe

If the video player tries to play an unknown type file, it will enter
the error_cb() function, which is called in another thread.
Use gdk_threads_enter/gdk_threads_leave to ensure the safe of gtk
multi-thread operation.

This fixes [BUGID #474]

Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
13 years agolibffi: bumpt to version 3.0.9
Joshua Lock [Fri, 15 Oct 2010 07:30:46 +0000 (08:30 +0100)]
libffi: bumpt to version 3.0.9

3.0.8 is incompatible with the autotools versions we use in Poky

Signed-off-by: Joshua Lock <josh@linux.intel.com>
13 years agomake: enable make-native
Joshua Lock [Fri, 15 Oct 2010 07:27:11 +0000 (08:27 +0100)]
make: enable make-native

Fedora 14 ships with a version of make which doesn't work with the Makefiles
shipped in our busybox and eglibc versions. Enable make native as a workaround.

This is a low risk fix for [BUGID #467]

Signed-off-by: Joshua Lock <josh@linux.intel.com>
13 years agohandbook: Fix the Ubuntu version requirement
Scott Rifenbark [Thu, 30 Sep 2010 18:36:26 +0000 (11:36 -0700)]
handbook: Fix the Ubuntu version requirement

Signed-off-by: Scott Rifenbark <scott.m.rifenbark@intel.com>
13 years agohandbook: Fix section numbering
Scott Rifenbark [Thu, 30 Sep 2010 18:36:26 +0000 (11:36 -0700)]
handbook: Fix section numbering

Signed-off-by: Scott Rifenbark <scott.m.rifenbark@intel.com>
13 years agoatom-pc: remove kernel video mode parameters
Darren Hart [Thu, 14 Oct 2010 19:33:56 +0000 (12:33 -0700)]
atom-pc: remove kernel video mode parameters

The kernel video mode parameters are not needed with the atom-pc linux-wrs
configuration. Without them, we boot up with a standard 80x25 text display
until the Intel i915 driver is initialized, at which point we switch to a
higher resolution text mode. Afterwards, a sato image will start the splash
screen and successfully boot to X.

Tested on the Toshiba NB305 netbook and the Black Sand development board.

[BUGID #450]

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
CC: Tom Zanussi <tom.zanussi@intel.com>
CC: Bruce Ashfield <bruce.ashfield@windriver.com>
CC: Richard Purdie <rpurdie@linux.intel.com>
13 years agolinux-wrs: update SRCREV for mpc8315e-rdb
Bruce Ashfield [Thu, 14 Oct 2010 20:47:02 +0000 (16:47 -0400)]
linux-wrs: update SRCREV for mpc8315e-rdb

Fixes [BUGID: 423]

The mpc8315e-rdb BSP failed to build due to -Os being used
in the boot code. The fix that doesn't involve changing
the compiler is to remove -Os in the powerpc code and keep
it contained on the BSP branch.

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
13 years agogupnp: fix broken SRC_URIs
Scott Garman [Thu, 14 Oct 2010 20:45:27 +0000 (13:45 -0700)]
gupnp: fix broken SRC_URIs

This fixes [BUGID #469]

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
13 years agoemenlow: add 8250 serial options
Bruce Ashfield [Thu, 14 Oct 2010 18:55:51 +0000 (14:55 -0400)]
emenlow: add 8250 serial options

The emenlow didn't have the default 8250 serial ports.
Including the common serial config fragment fixes this
problem.

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
13 years agopoky-qemu-internal: make the -no-reboot qemu option mandatory
Scott Garman [Thu, 14 Oct 2010 18:42:13 +0000 (11:42 -0700)]
poky-qemu-internal: make the -no-reboot qemu option mandatory

This fixes [BUGID #100]

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
13 years agopoky-qemu: miscellaneous bugfixes
Scott Garman [Thu, 14 Oct 2010 18:29:44 +0000 (11:29 -0700)]
poky-qemu: miscellaneous bugfixes

In scenarios where the POKY_NATIVE_SYSROOT env variable hasn't been
set up, bug #427 can still be triggered. This fixes it by running
setup_tmpdir(). This fixes [BUGID #427].

Also, the qemu tap dev lock directory needs to be chmod 777 so that
multiple users can create and delete lock files.

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
13 years agopoky-qemu: add nographic option for disabling video console
Scott Garman [Thu, 14 Oct 2010 18:23:26 +0000 (11:23 -0700)]
poky-qemu: add nographic option for disabling video console

This fixes [BUGID #408]

Signed-off-by: Scott Garman <scott.a.garman@intel.com>
13 years agoatom-pc: get rid of undefined video mode prompt on bootup
Tom Zanussi [Thu, 14 Oct 2010 18:38:47 +0000 (13:38 -0500)]
atom-pc: get rid of undefined video mode prompt on bootup

Temporary fix for [BUGID #450]

This fixes the video mode number to the expected value for this video
setting, but the setting itself needs to be fixed along with matching
config options - this patch is just a temporary fix for RC2.

Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>
13 years agosysvinit: add check for no X11 to disbale psplash
Saul Wold [Thu, 14 Oct 2010 16:48:21 +0000 (09:48 -0700)]
sysvinit: add check for no X11 to disbale psplash

Add a check to the rc script to check if the /etc/init.d/xserver-nodm
script exists and is executable and not disable psplash if it is,
otherwise disable pspalsh since we do not have X installed.

Fixed [BUG #457]

Signed-off-by: Saul Wold <Saul.Wold@intel.com>
13 years agotask-poky-qt: Re-Enable MIPS Build and disable fotowall
Saul Wold [Thu, 14 Oct 2010 15:46:01 +0000 (08:46 -0700)]
task-poky-qt: Re-Enable MIPS Build and disable fotowall

This address a number if issues with fotowall not being ready
and [BUGID #462] Qt missing for Mips build

Signed-off-by: Saul Wold <Saul.Wold@intel.com>
13 years agobitbake-runtask: Ensure logging settings take effect in the worker process (and pass...
Richard Purdie [Tue, 12 Oct 2010 16:46:47 +0000 (17:46 +0100)]
bitbake-runtask: Ensure logging settings take effect in the worker process (and pass verbose setting)

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agopseudo: Update to revision 489ed5e89f040febcb605f22187968f597211205
Richard Purdie [Thu, 14 Oct 2010 19:03:58 +0000 (20:03 +0100)]
pseudo: Update to revision 489ed5e89f040febcb605f22187968f597211205

This fixes various issues with modern coreutils which use *xattr() syscalls
which pseudo wasn't catching causing permissions problems.

[BUGID #463]

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agoeglibc-nativesdk: Add patch to change the search path of the dynamic linker
Richard Purdie [Thu, 14 Oct 2010 10:44:57 +0000 (11:44 +0100)]
eglibc-nativesdk: Add patch to change the search path of the dynamic linker

Change the search path of the dynamic linker to search ld.so.cache as a last resort when
trying to find libraries at runtime. This means libs in /opt/poky are used in preference
but any host libs such as libGL can also be found. See the patch for a more detailed
description.

Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
13 years agoRevert "meta-toolchain: Update the way ld.so.cache is handled"
Richard Purdie [Thu, 14 Oct 2010 10:42:37 +0000 (11:42 +0100)]
Revert "meta-toolchain: Update the way ld.so.cache is handled"

This change didn't work as expected as ld.so.cache doesn't preseve any search
path ordering, therefore revert the change.

This reverts commit b2c3ce6d17520d43b10777dd306449876cb40819.

13 years agolinux-wrs: update SRCREVs
Bruce Ashfield [Thu, 14 Oct 2010 00:15:31 +0000 (20:15 -0400)]
linux-wrs: update SRCREVs

Fixes [BUGID #432, #438, #437, #422]

    [atom-pc: switch to the atom-pc branch

    The atom-pc has a dedicated branch and merged kernel commits.
    Swtich the kernel recipe to use the fully integrated BSP

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>]
    [atom-pc: add kernel config options needed for -live images

    Fixes [BUGID #432]

    Live images require VFAT filesystem and loop device support - this
    adds the required kernel config options for them.

Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>]
    [atom-pc: Add netbook wireless networking support for NB305 and eee901

    Add the atom-pc-wifi.cfg to enable wireless networking support for the netbooks
    mentioned in the atom-pc machine config file.

    Note: this adds a staging tree driver (rt2860sta) for the eee901 wifi.

    [BUGID #438]

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
CC: Tom Zanussi <tom.zanussi@intel.com>
CC: Richard Purdie <rpurdie@linux.intel.com>
CC: Bruce Ashfield <bruce.ashfield@windriver.com>]
    [atom-pc: Add netbook wired networking support for NB305, Aspire One, and eee901

    Add the atom-pc-eth.cfg to enable wired networking support for the three
    netbooks mentioned in the atom-pc machine config file.

    [BUGID #437]

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
CC: Tom Zanussi <tom.zanussi@intel.com>
CC: Richard Purdie <rpurdie@linux.intel.com>
CC: Bruce Ashfield <bruce.ashfield@windriver.com>]
    [routerstationpro: add missing kernel patches

    [BUGID: #422]

    Add the kernel patches required for routerstaion pro boot

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>]
    [netbook: allow hardisk booting

    [BUGID: #445]

Signed-off-by: Darren Hart <dvhart@linux.intel.com>]
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
13 years agoatom-pc: switch to the atom-pc branch
Bruce Ashfield [Wed, 13 Oct 2010 20:27:27 +0000 (16:27 -0400)]
atom-pc: switch to the atom-pc branch

The atom-pc has a dedicated branch and merged kernel commits.
Swtich the kernel recipe to use the fully integrated BSP

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>