| Age | Commit message (Collapse) | Author |
|---|
|
git-svn-id: https://svn.o-hand.com/repos/poky/trunk@5116 311d38ba-8fff-0310-9ca6-ca027cbcb966
|
|
|
|
|
|
|
|
git-svn-id: https://svn.o-hand.com/repos/poky/trunk@5388 311d38ba-8fff-0310-9ca6-ca027cbcb966
|
|
git-svn-id: https://svn.o-hand.com/repos/poky@5389 311d38ba-8fff-0310-9ca6-ca027cbcb966
|
|
git-svn-id: https://svn.o-hand.com/repos/poky@5382 311d38ba-8fff-0310-9ca6-ca027cbcb966
|
|
Squashed changes from Poky:
svn r5067 - package_rpm.bbclass: Update against recent packaging changes
and start to implement dependency handling.
svn r5100 - package_rpm.bbclass: Fix empty package handling, handle
package dependency renaming by adding a missing function
call and expand RPMBUILD before changing the dictonary so
local PV/PR changes don't break the WORKDIR variable
expansion
svn r5111 - package_rpm.bbclass: Stop autogenerated dependencies for
now, its not functioning correctly
svn r5116 - package_rpm.bbclass: files needs to start from /
svn r5129 - package_rpm.bbclass: Fix problems with package version with
'-' in the version (replace with '+'), improve the Requires
and Recommends generation code
svn r5131 - package_rpm.bbclass: Add support for postinst and preinst scripts
svn r5381 - classes: Split rpm feeds by PACKAGE_ARCH
svn r5390 - package_rpm.bbclass: Improve handling of '-' characters in Requires
and Recommends fields
svn r5393 - package_rpm.bbclass: Fix package version fixing for PKG renamed packages
svn r5397 - package_rpm.bbclass: Fix dependency rebuilding bug
svn r5413 - package_rpm.bbclass: save value of RPMBUILDPATH as it is needed for
packaging
package_rpm.bbclass: Don't use the system wide RPM macros.
This fixes the build on Fedora 10 where the RPM system macros obliterate
the value of BuildRoot that is set in the spec file.
package_rpm.bbclass: fix indentation package_rpm.bbclass: fix syntax errors
package_rpm.bbclass: save RPROVIDES in packages
|
|
Squashed changes from Poky:
svn r5101 - Add rootfs_rpm.bbclass, a start at image generation from rpms. Its
not expected to be entirely functional yet.
svn r5112 - rootfs_rpm.bbclass: Remove the feed directory before packaging the image
svn r5134 - rootfs_rpm.bbclass: Add basic postinstall handling during image creation
svn r5143 - rootfs_rpm.bbclass: Fixup broken symlinks due to fakechroot and run
any leftover postinstalls on real hardware at first boot
svn r5158 - rootfs_rpm.bbclass: Add recommends handling
svn r5380 - rootfs_rpm.bbclass: ship rpm and yum in target images
svn r5382 - rootfs_rpm.bbclass: Add rpm package coverage
svn r5386 - rootfs_rpm.bbclass: use more efficient way to copy packages
svn r5391 - rootfs_rpm: set target architecture and do not tell yum which arch we will use
svn r5398 - rootfs_rpm.bbclass: Set YUM_ARCH_FORCE
svn r5412 - rootfs_rpm.bbclass: remove lock files to get rpm working on device
svn r5432 - rootfs_rpm.bbclass: fix platform so rpm will be usable on device
svn r5433 - rootfs_rpm.bbclass: remove temporary yum config - rootfs has one
from yum package
svn r5434 - rootfs_rpm.bbclass: added generation of feed entries
svn r5436 - rootfs_rpm.bbclass: added code to generate images from feeds
OE changes:
rootfs_rpm.bbclass: remove two functions (need rewrite for RPM)
install_all_locales and remove_packaging_data_files functions were from
rootfs_ipk.bbclass and need to be rewritten for RPM rootfs.
rootfs_rpm.bbclass: use DISTRO instead of hardcoded Poky - spotted by Tom Rini
rootfs_rpm.bbclass: drop not needed disabled code
|
|
|
|
to packages/ while parsing BBFILES
Signed-off-by: Denys Dmytriyenko <denis@denix.org>
Acked-by: Mike Westerhof <mwester@dls.net>
Acked-by: Philip Balister <philip@balister.org>
Acked-by: Khem Raj <raj.khem@gmail.com>
Acked-by: Marcin Juszkiewicz <hrw@openembedded.org>
Acked-by: Koen Kooi <koen@openembedded.org>
Acked-by: Frans Meulenbroeks <fransmeulenbroeks@gmail.com>
|
|
* remove qtopia-core package, superseeded by qt-embedded
* remove qtopia4core class, superseed by qt4e
Acked-By: Mickey Lauer <mickey@vanille-media.de>
|
|
vala.bbclass: new class for vala-based packages
|
|
into org.openembedded.dev
|
|
Juszkiewicz <hrw@openembedded.org>)
- converted tabs to spaces
- allow entries which have no corresponding jar file (just skip)
- document that behavior
|
|
into org.openembedded.dev
|
|
tempfile.mkstemp() is used to create a temporary file. This function
return a tuple with an OS file descriptor and a filename. Filename is
stored in "tmpfile" but descriptor is not stored anywhere, but it is
still open because it's only an integer to python so it is not closed at
the end of the function.
For each iteration in which this function is called, a new OS file
descriptor is opened, but not closed. The solution is to store the file
descriptor and close it.
|
|
* Current pattern match does one think wrong consider
something like this
dependency_libs=' -L/home/kraj/work/oe/build/eglibc/omap5912osk/work/x86_64-linux/ncurses-native-5.4-r17/ncurses-5.4/lib'
this pattern when matched gets converted to
dependency_libs=' -L/home/kraj/work/oe/build/eglibc/omap5912osk/staging/x86_64-linux/usr/lib/lib'
which is wrong. If we had a '/' appended to original path then it would work.
This patch lets it straight.
|
|
This reverts commit 25b281668e62d90c027978b13b53075534b81d0c.
As pointed out by Pieter <pieterg@gmx.com> my logic here was backwards
and the code was correct before and this change is incorrect.
|
|
|
|
|
|
It was removing file by file before and did not removed conffiles file.
As a result package_rpm was not able to create package.
|
|
Conflicts:
classes/qt4e.bbclass
packages/bluez/bluez4_4.31.bb
packages/mc/mc_4.6.2.bb
packages/vte/vte.inc
packages/vte/vte_0.16.13.bb
packages/vte/vte_0.16.14.bb
|
|
|
|
|
|
|
|
|
|
|
|
|
|
rootfs_ipk: task-boot change fixes
|
|
images as well
* This fixes the "image broken due to not running postinsts because opkg* isn't installed" type of bugs
|
|
update users as well
Acked-By: Marcin Juszkiewicz <hrw@openembedded.org>
|
|
before (with BitBake trunk):
NOTE: package linux-2.6.28-r4: task do_compile: started
NOTE: package linux-2.6.28-r4: task do_compile: Started
after:
NOTE: package linux-2.6.28-r4: task do_compile: Started
|
|
build
|
|
|
|
|
|
git://projetos.ossystems.com.br/git/users/otavio/org.openembedded.dev into org.openembedded.dev
* 'master' of git://projetos.ossystems.com.br/git/users/otavio/org.openembedded.dev:
flash-plugin: add 10.0.12.36
sun-jre6: add 1.6.0.10
parted: drop versions up to 1.8.8
package.bbclass, package_ipk.bbclass: add dynamic depends on locale packages (ipk specific)
cups: add 1.3.8.
rt61-firmware: add 1.2
|
|
(ipk specific)
|
|
|
|
With the previously broken and unused KERNEL_OBJECT_SUFFIX, this actually
evaluted to "" so we always did a find path -name *, which is why we need
the -d test later on. Switch to testing for KERNEL_OBJECT_SUFFIX==2.6
and else'ing for 2.4 and doing .ko or .o. Leave KERNEL_OBJECT_SUFFIX in-tact
for non-mainline users. This was easier than getting kernel.bbclass to
evaluate KERNEL_OBJECT_SUFFIX either on its own or re-jigging module-base and
module_strip, again for non-mainline users.
|
|
We were saying if we're newer than 2.6.0, use ".o" and otherwise use ".ko",
which is backwards.
|
|
|
|
MIPS doesn't do GNU_HASH so if objdump tells us "[mips32]" or "[mips64]"
then set sane = True
|
|
gtk+ fails with dash
|
|
is set * we want to error out on non-matching sums, missing ones are less bad
* People can now set OE_STRICT_CHECKSUMS = "" in local.conf to make a missing checksum a warning, but checksum errors are still fatal, as it should
|
|
|
|
|
|
classes that make sure correct files end up on rootfs
|
|
postprocessing on it with do_configure_append
|
|
|
Marcin Juszkiewicz
Robert Schuster
Richard Purdie
Denys Dmytriyenko
Jeremy Lainé
Michael 'Mickey' Lauer
Khem Raj
Robert Schuster
Jader H. Silva
Tom Rini
Marcin Juszkiewicz
Koen Kooi
Roman I Khimov
Otavio Salvador
woglinde
Graeme Gregory