Hello, On Sun, Jan 26, 2014 at 4:06 PM, Philip Balister <[email protected]> wrote: > On 01/26/2014 12:05 PM, Koen Kooi wrote: >> >> Op 26 jan. 2014, om 11:11 heeft Robert Yang <[email protected]> het >> volgende geschreven: >> >>> >>> >>> On 01/26/2014 05:32 PM, Koen Kooi wrote: >>>> >>>> Op 26 jan. 2014, om 07:16 heeft Robert Yang <[email protected]> >>>> het volgende geschreven: >>>> >>>>> Note: >>>>> This PULL is based on oe-core, and here is another one based on poky: >>>>> >>>>> git://git.yoctoproject.org/poky-contrib robert/dora-next >>>>> >>>>> // Robert >>>>> >>>>> >>>>> The following changes since commit >>>>> 198623d80d31f19c963e61d03cbcb12dd318dfdf: >>>>> >>>>> ltp: set PREFERRED_PROVIDER and rename runtests_noltp.sh script >>>>> (2014-01-24 12:48:15 +0000) >>>>> >>>>> are available in the git repository at: >>>>> >>>>> git://git.openembedded.org/openembedded-core-contrib robert/dora-next >>>>> >>>>> http://cgit.openembedded.org/cgit.cgi/openembedded-core-contrib/log/?h=robert/dora-next >>>>> >>>>> Alexandre Belloni (1): >>>>> wpa-supplicant-2.0: don't exit in pkg_postinst >>>>> >>>>> Anders Darander (1): >>>>> terminal.bbclass: do not export PS1 >>>>> >>>>> Andreas Oberritter (1): >>>>> cogl-1.0: explicitly disable cairo >>>>> >>>>> Bruce Ashfield (1): >>>>> linux-libc-headers: fix MIPS klibc build error >>>>> >>>>> Chen Qi (1): >>>>> subversion: fix build problem when sysroot contains '-D' or '-I' >>>>> >>>>> Christopher Larson (4): >>>>> python, python-native: fix PARALLEL_MAKEINST failure >>>>> cairo: add/use packageconfig for valgrind support >>>>> pulseaudio: fix RDEPENDS traversal for consolekit >>>>> xz: make the LICENSE info more accurate >>>>> >>>>> Cristiana Voicu (1): >>>>> babeltrace: correct PV variable >>>>> >>>>> Enrico Scholz (1): >>>>> bluez4: added dependency on 'libsndfile1' >>>>> >>>>> Hongxu Jia (5): >>>>> kconfig-frontends: add python to kconfig-frontends's RDEPENDS >>>>> qemu: add PACKAGECONFIG for vnc, libcurl, nss, uuid, curses, gtk+, >>>>> libcap-ng >>>>> qemu: add bash and python to qemu's RDEPENDS >>>>> license.bbclass: fix copying license directories failed >>>>> nativesdk.bbclass: support nativesdk to override with the >>>>> PACKAGES_DYNAMIC statement >>>>> >>>>> Jackie Huang (3): >>>>> grub: add explicit dependency on bison-native >>>>> kconfig-frontends: fix the incorrect depends on gperf >>>>> guile: fix the depends for target recipes >>>>> >>>>> Jacob Kroon (1): >>>>> meta/lib/oe/terminal.py: Don't pass non-supported '--disable-factory' >>>>> flag to gnome-terminal >>>>> >>>>> Kai Kang (1): >>>>> x264: install libraries to right directory when enable multilib >>>>> >>>>> Krzysztof Sywula (1): >>>>> Minicom depends on libiconv >>>>> >>>>> Laurentiu Palcu (1): >>>>> x11vnc: fix CAPS_LOCK issues >>>>> >>>>> Li Wang (1): >>>>> xinetd: CVE-2013-4342 >>>>> >>>>> Martin Jansa (3): >>>>> avahi: add leading space to RRECOMMENDS append >>>>> xinput-calibrator: add formfactor to RDEPENDS >>>>> base.bbclass: Set umask 022 also for do_unpack task >>>>> >>>>> Ming Liu (4): >>>>> qemu: use PACKAGECONFIG to address xfsprogs dependency >>>>> qemu: explicitly disable xen support >>>>> libpthread-stubs: should set ALLOW_EMPTY >>>>> grub: add PACKAGECONFIG for device-mapper >>>>> >>>>> Nick D'Ademo (1): >>>>> libav: install libraries to right directory when multilib is enabled >>>>> >>>>> Otavio Salvador (1): >>>>> gcc-4.8: Backport PR c++/57532 fix from 4.8.2 >>>>> >>>>> Paul Eggleton (2): >>>>> linux-firmware: add missing linux-firmware-iwlwifi-7260-7 package >>>>> libav: add libpostproc to PROVIDES (for 0.8.x version only) >>>>> >>>>> Phil Blundell (2): >>>>> libsoup: Remove libproxy from DEPENDS >>>>> binutils: Also add autoconf-native to DEPENDS >>>>> >>>>> Richard Purdie (7): >>>>> base/gcc-common: Ensure umask setting is consistent for shared workdir >>>>> image.bbclass: Depend on virtual/kernel:do_deploy >>>>> gcc-cross-canadian: Fix fortran build >>>>> gcc-crosssdk.inc: Fix missing dependencies (such as libmpc-native) >>>>> eglibc-locale: Fix depends on binutils >>>>> eglibc-locale: Fix previous dependency change to properly work in >>>>> nativesdk case >>>>> eglibc-locale: Fix multilib builds to depend on the correct binutils >>>>> >>>>> Robert Yang (2): >>>>> coreutils 6.9: fix coreutils.texi >>>>> gcc-4.8/libstdc++-v3: disable sdt >>>>> >>>>> Ross Burton (3): >>>>> ptest: ensure do_install_ptest_base task runs in fakeroot context >>>>> useradd.bbclass: add dependency on base-files >>>>> dbus: use PACKAGECONFIG for X11 and systemd >>>>> >>>>> Roy Li (1): >>>>> multilib: Ensure we map the SYSTEMD_PACKAGES variable >>>>> >>>>> Saul Wold (1): >>>>> openssl: use PACKAGECONFIG to disable perl bits >>>>> >>>>> Scott Garman (1): >>>>> runqemu: remove core-image-* whitelist >>>>> >>>>> Tom Zanussi (1): >>>>> systemtap: Add --enable-prologues to configuration >>>>> >>>>> Yue Tao (3): >>>>> python: do not replace ccache in the middle of a path >>>>> acpid: CVE-2011-1159 >>>>> icu: CVE-2013-2924 >>>>> >>>>> mykhani (1): >>>>> openssl.inc: Install c_rehash utility with openssl >>>>> >>>>> yzhu1 (1): >>>>> populate_sdk: verify executable or dynamically linked library >>>> >>>> I absolutely hate these massive pull requests for stable branches: >>>> >>>> 1) important fixes keep collecting dust >>>> 2) 56 changes at once makes testing and debugging really painfull >>>> >>>> So for the future could you *please* send pull requests directly after >>>> patches get proposed? >>>> >>> >>> Sorry, I don't quite understand what do you mean, this is just for review, >>> not the final PULl, I only sent the cover lettter is for avoiding flushing >>> mailing list. >> >> So you're saying the final pull request is less massive than this? You still >> haven't explained why you're doing the big bang style of maintenance. > > Koen would like to smaller more frequent, smaller numbers of patches > applied. > > What is the current maintenance approach? > > 1) Updates performed at constant time intervals, regardless of patches. > 2) Updates performed when patch backlog hits N. > 3) Updates performed when maintainer has free time to review the patches.
I second Koen concern. It is not good to hold for too long and have a massive set of stable updates when we have people proposing patches for merging. I understand it requires testing but in case of regressions it is a nightmare to figure out the culprit. -- Otavio Salvador O.S. Systems http://www.ossystems.com.br http://code.ossystems.com.br Mobile: +55 (53) 9981-7854 Mobile: +1 (347) 903-9750 _______________________________________________ Openembedded-core mailing list [email protected] http://lists.openembedded.org/mailman/listinfo/openembedded-core
