Re: [oe] [meta-python][PATCH] python3-pycryptodomex: add 3.9.4 recipe

2020-03-11 Thread Tim Orling
On Wed, Mar 11, 2020 at 8:02 PM Scott Branden wrote: > So no changes to make. Could we get this added to master-next? > My concerns have been addressed. Thank you everyone for the time you took to research the state of things. Reviewed-by: Tim Orling > > On Tue, 10 Mar 2020 at 20:48,

Re: [oe] [meta-python][PATCH] python3-pycryptodomex: add 3.9.4 recipe

2020-03-11 Thread Scott Branden via Openembedded-devel
So no changes to make. Could we get this added to master-next? On Tue, 10 Mar 2020 at 20:48, Rajesh Ravi wrote: > ++ > > Hi Scott, > > A)I tried a lot no other recipe could resolve OP-TEE 3.8.0 build break. > I posted this in OP-TEE as well as meta-openembedded forums. > This is the only

[oe] [zeus 11/13] opencv: abort configure if we need to download

2020-03-11 Thread Armin Kuster
From: Ross Burton OpenCV's habit of downloading files during do_configure is bad form (as it becomes impossible to do offline builds), so add an option to error out if a download would be needed. Signed-off-by: Ross Burton Signed-off-by: Khem Raj Signed-off-by: Yeoh Ee Peng Signed-off-by:

[oe] [zeus 13/13] sanlock: Replace cp -a with cp -R --no-dereference

2020-03-11 Thread Armin Kuster
From: Khem Raj helps to stop leaking builder's UID into sstate cache Fixes Exception: KeyError: 'getpwuid(): uid not found: 6000' Signed-off-by: Khem Raj (cherry picked from commit 80d4d7538a85123603eea3d87542c1c0433febb7) [ak: fixes build issue Exception: KeyError: 'getpwuid(): uid not

[oe] [zeus 12/13] s3c24xx-gpio, s3c64xx-gpio, sjf2410-linux-native, usbpath, wmiconfig: use git fetcher instead of svn fetcher

2020-03-11 Thread Armin Kuster
From: Martin Jansa * svn checkouts from http://svn.openmoko.org/ are now redirected to github: svn --non-interactive --trust-server-cert co --no-auth-cache --ignore-externals -r 4949 http://svn.openmoko.org/trunk/src/target/gpio@4949 gpio Redirecting to URL

[oe] [zeus 09/13] opencv: also download face alignment data in do_fetch()

2020-03-11 Thread Armin Kuster
From: Ross Burton The face alignment data is downloaded in do_configure, so download it in do_fetch and add it to the cache. Signed-off-by: Ross Burton Signed-off-by: Khem Raj Signed-off-by: Yeoh Ee Peng Signed-off-by: Armin Kuster --- meta-oe/recipes-support/opencv/opencv_4.1.0.bb | 3 +++

[oe] [zeus 10/13] opencv: PACKAGECONFIG for G-API, use system ADE

2020-03-11 Thread Armin Kuster
From: Ross Burton The Graph API is enabled by default, and if ADE isn't present it will download a copy of the source during do_configure. Add a PACKAGECONFIG for the Graph API, and depend on the ADE that we package. Signed-off-by: Ross Burton Signed-off-by: Khem Raj Signed-off-by: Yeoh Ee

[oe] [zeus 02/13] wireshark: Upgrade 3.0.6 -> 3.0.8

2020-03-11 Thread Armin Kuster
From: Adrian Bunk Upgrade on the 3.0 stable branch, including fixes for CVE-2019-19553 and CVE-2020-7045. Signed-off-by: Adrian Bunk Signed-off-by: Armin Kuster --- .../wireshark/{wireshark_3.0.6.bb => wireshark_3.0.8.bb} | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) rename

[oe] [zeus 07/13] opencv: Enable pkg-config .pc file generation

2020-03-11 Thread Armin Kuster
From: Carlos Rafael Giani In OpenCV 4, .pc file generation is disabled by default. Yet, other software such as GStreamer and FFmpeg rely on the .pc files during build time configuration. Explicitely enable .pc file generation to make sure pkg-config can be used for getting information about

[oe] [zeus 08/13] opencv: don't download during configure

2020-03-11 Thread Armin Kuster
From: Ross Burton OpenCV downloads data files during the CMake configure phase, which is bad because fetching should only happen in do_fetch (and if proxies are needed, won't be set in do_configure). The recipe attempts to solve this already by having the repositories in SRC_URI and moving the

[oe] [zeus 05/13] libldb: Do not require the "pam" distro feature to be enabled

2020-03-11 Thread Armin Kuster
From: Peter Kjellerstedt It was only added because samba was a dependency, but was not removed again when the dependency on samba was removed in commit 6207331f. This effectively reverts commit a190c2e3. Signed-off-by: Peter Kjellerstedt Signed-off-by: Khem Raj (cherry picked from commit

[oe] [zeus 06/13] ade: Fix install paths in multilib builds

2020-03-11 Thread Armin Kuster
From: Khem Raj Fixes ERROR: ade-0.1.1f-r0 do_package: QA Issue: ade: Files/directories were installed but not shipped in any package: /usr/lib /usr/lib/libade.a Signed-off-by: Khem Raj Signed-off-by: Sanjeev Nahulanthran Signed-off-by: Armin Kuster ---

[oe] [zeus 01/13] mosquitto: Use mosquitto.init for daemon init

2020-03-11 Thread Armin Kuster
From: Mike Krupicka Config file specification is missing in start) case. It is present already in restart) case. Signed-off-by: Khem Raj (cherry picked from commit 257ea010b716073acbe6866f4c3968b4962fdc37) Signed-off-by: Armin Kuster ---

[oe] [zeus 04/13] lmsensors: Fix sensord dependencies

2020-03-11 Thread Armin Kuster
From: Paul Barker If sensord is removed from PACKAGECONFIG, the recipe should not depend on rrdtool and the lmsensors package should not depend on lmsensors-sensord. Signed-off-by: Paul Barker Signed-off-by: Khem Raj Backport of 5674b0a9e8778f7538419451e629ef0c13711123 from master branch.

[oe] [zeus 03/13] lvm2, libdevmapper: Do not patch configure

2020-03-11 Thread Armin Kuster
From: Peter Kjellerstedt Since the do_configure task regenerates the configure script, there is no need to patch it. Actually doing so will cause problems, which can be seen by doing: bitbake lvm2 -c configure bitbake lvm2 -c patch -f Reported-by: Andrey Zhizhikin Signed-off-by: Peter

[oe] [zeus 00/13] Patch review

2020-03-11 Thread Armin Kuster
Here is the next set for meta-openembedded. Please have reviews back by Friday. The following changes since commit bb65c27a772723dfe2c15b5e1b27bcc1a1ed884c: fluentbit: Fix packaging in multilib env (2020-01-30 18:38:10 -0800) are available in the Git repository at:

Re: [oe] State of OE world - 2020-03-09

2020-03-11 Thread Khem Raj
On Wed, Mar 11, 2020 at 4:12 PM akuster808 wrote: > > Khem, > > There is something wrong with this. > > No one will be able to get to those logs posted in this email or what is on > the wiki. > > I think we need some links like: > >

Re: [oe] State of OE world - 2020-03-09

2020-03-11 Thread akuster808
Khem, There is something wrong with this. No one will be able to get to those logs posted in this email or what is on the wiki. I think we need some links like: http://logs.nslu2-linux.org/buildlogs/oe/world/dunfell/log.world.qemuarm.20191213_215925.log/ -armin On 3/11/20 3:40 PM, Khem Raj

Re: [oe] [meta-ota][PATCH] meta-ota: add support for binary-delta images in a new layer

2020-03-11 Thread Otavio Salvador
On Wed, Mar 11, 2020 at 10:28 AM Nicolas Dechesne wrote: > On Wed, Mar 11, 2020 at 9:04 AM Bartosz Golaszewski wrote: >> wt., 3 mar 2020 o 14:56 Bartosz Golaszewski napisał(a): >> > >> > If this is not something that should be part of meta-openembedded - is >> > there any way to have it hosted

[oe] State of OE world - 2020-03-09

2020-03-11 Thread Khem Raj
== Failed tasks 2020-03-09 == INFO: jenkins-job.sh-1.8.46 Complete log available at /media/ra_build_share/buildlogs/oe/world/dunfell/log.report.20200310_140304.log === common (0) === === common-x86 (0) === === qemuarm (0) === === qemuarm64 (1) === *

Re: [oe] [meta-oe][PATCH v2] openjpeg: Fix CVE-2020-6851

2020-03-11 Thread akuster808
On 3/11/20 12:11 PM, Sakib Sajal wrote: > From: Yue Tao > > Backport patch from upstream to fix heap-based buffer overflow > > Upstream-Status: Backport > CVE: CVE-2020-6851 Applies to Zeus too. > > Signed-off-by: Yue Tao > Signed-off-by: Sakib Sajal > --- >

[oe] [meta-oe][PATCH v2] openjpeg: Fix CVE-2020-6851

2020-03-11 Thread Sakib Sajal
From: Yue Tao Backport patch from upstream to fix heap-based buffer overflow Upstream-Status: Backport CVE: CVE-2020-6851 Signed-off-by: Yue Tao Signed-off-by: Sakib Sajal --- .../openjpeg/openjpeg/CVE-2020-6851.patch | 32 +++ .../openjpeg/openjpeg_2.3.1.bb

Re: [oe] [meta-python][PATCH 1/3] python3-cycler: add recipe for 0.10.0

2020-03-11 Thread Tim Orling
FWIW, I had made significant progress towards getting ptest enabled for Matplotlib, although a bit out of date now [1]. Which included a recipe for cycler. It also unbundled a bunch of dependencies, like libagg. The ptests do require a fair amount of ram, so standard qemu settings will epically

Re: [oe] [meta-oe][PATCH] layer.conf: add meta-python to LAYERDEPENDS

2020-03-11 Thread Tim Orling
On Wed, Mar 11, 2020 at 12:38 AM Khem Raj wrote: > On Tue, Mar 10, 2020 at 11:39 PM Changqing Li > wrote: > > > > > > On 3/7/20 12:12 AM, Khem Raj wrote: > > > On Thu, Mar 5, 2020 at 11:55 PM Martin Jansa > wrote: > > >> Can we fix mozjs instead? > > > perhaps thats better fix, maybe it can

Re: [oe] [meta-ota][PATCH] meta-ota: add support for binary-delta images in a new layer

2020-03-11 Thread Nicolas Dechesne
hi Bartosz, On Wed, Mar 11, 2020 at 9:04 AM Bartosz Golaszewski wrote: > wt., 3 mar 2020 o 14:56 Bartosz Golaszewski napisał(a): > > > > If this is not something that should be part of meta-openembedded - is > > there any way to have it hosted at https://git.yoctoproject.org/cgit/ > > as an

Re: [oe] [meta-oe][PATCH] smem: matplotlib is an optional depencency

2020-03-11 Thread nick83ola
Hi Andre, the problem is that if you put the package in RRECOMMENDS, yocto will still build python3-numpy and python3-matplotlib even if you add it to BAD_RECOMMENDATIONS from https://www.yoctoproject.org/docs/current/mega-manual/mega-manual.html#var-RRECOMMENDS The package manager will

Re: [oe] [meta-oe][PATCH] smem: matplotlib is an optional depencency

2020-03-11 Thread André Draszik
On Wed, 2020-03-11 at 09:43 +, André Draszik wrote: > On Wed, 2020-03-11 at 07:55 +, Nicola Lunghi wrote: > > smem doesn't need matplotlib and numpy to run, is an optional dependency > > only > > needed to produce graphs. > > Right, that's why it's an RRECOMMENDS. If you don't want it in

[oe] [meta-oe][zeus][PATCH] rng-tools: Supply patches to fix rngd block reboot issue

2020-03-11 Thread De Huo
There will be issue about "Hardware RNG Entropy Gatherer Daemon slow stop" during "reboot" of CN96xx. Backport the patches from https://github.com/nhorman/rng-tools/commit/3e47faae108df4824531bf9c003cc1c65b7b2842 and

Re: [oe] [meta-oe][PATCH] smem: matplotlib is an optional depencency

2020-03-11 Thread André Draszik
On Wed, 2020-03-11 at 07:55 +, Nicola Lunghi wrote: > smem doesn't need matplotlib and numpy to run, is an optional dependency only > needed to produce graphs. Right, that's why it's an RRECOMMENDS. If you don't want it in the image, just do BAD_RECOMMENDATIONS as per usual practice :-)

[oe] [meta-oe][PATCH] conf/layer.conf: add BBFILES_DYNAMIC and dynamic layers

2020-03-11 Thread changqing.li
From: Changqing Li some recipes under meta-oe have dependency on meta-python, and test_world of yocto-check-layer will failed with error like: ERROR: test_world (common.CommonCheckLayer) ERROR: Nothing PROVIDES 'python3-pytoml-native' (but

Re: [oe] [meta-ota][PATCH] meta-ota: add support for binary-delta images in a new layer

2020-03-11 Thread Bartosz Golaszewski
wt., 3 mar 2020 o 14:56 Bartosz Golaszewski napisał(a): > > If this is not something that should be part of meta-openembedded - is > there any way to have it hosted at https://git.yoctoproject.org/cgit/ > as an official yocto layer? I'm sorry if this is a dumb question, I > don't know exactly

[oe] [meta-oe][PATCH] smem: matplotlib is an optional depencency

2020-03-11 Thread Nicola Lunghi
smem doesn't need matplotlib and numpy to run, is an optional dependency only needed to produce graphs. This is needed to avoid to include and build matplotlib and numpy by default Signed-off-by: Nicola Lunghi --- meta-oe/recipes-support/smem/smem_1.5.bb | 4 +++- 1 file changed, 3

Re: [oe] [meta-oe][PATCH] layer.conf: add meta-python to LAYERDEPENDS

2020-03-11 Thread Khem Raj
On Tue, Mar 10, 2020 at 11:39 PM Changqing Li wrote: > > > On 3/7/20 12:12 AM, Khem Raj wrote: > > On Thu, Mar 5, 2020 at 11:55 PM Martin Jansa wrote: > >> Can we fix mozjs instead? > > perhaps thats better fix, maybe it can made optional via packageconfig ? > > or marked incompatible if

[oe] [meta-python][PATCH 3/3] python3-matplotlib: fix dependencies and license

2020-03-11 Thread Nicola Lunghi
* The license indicated in the old recipe (and in this) is the PSF not the BSD that was indicated before. * fixed the dependencies (looking at the recipe at meta-jupiter https://github.com/Xilinx/meta-jupyter/tree/master/recipes-python/ * the setupext script was a leftover from the python2 recipe

[oe] [meta-python][PATCH 1/3] python3-cycler: add recipe for 0.10.0

2020-03-11 Thread Nicola Lunghi
This is a dependency for python3-matplotlib Signed-off-by: Nicola Lunghi --- .../python/python3-cycler_0.10.0.bb | 16 1 file changed, 16 insertions(+) create mode 100644 meta-python/recipes-devtools/python/python3-cycler_0.10.0.bb diff --git

[oe] [meta-python][PATCH 2/3] python3-kiwisolver: add recipe for version 1.1.0

2020-03-11 Thread Nicola Lunghi
This is a dependency for python3-matplotlib Signed-off-by: Nicola Lunghi --- .../python/python3-kiwisolver_1.1.0.bb | 16 1 file changed, 16 insertions(+) create mode 100644 meta-python/recipes-devtools/python/python3-kiwisolver_1.1.0.bb diff --git

Re: [oe] [meta-oe][PATCH] layer.conf: add meta-python to LAYERDEPENDS

2020-03-11 Thread Changqing Li
On 3/7/20 12:12 AM, Khem Raj wrote: On Thu, Mar 5, 2020 at 11:55 PM Martin Jansa wrote: Can we fix mozjs instead? perhaps thats better fix, maybe it can made optional via packageconfig ? or marked incompatible if meta-python is not included Hi,  Khem There are 2 solutions for the problem,