Upgrade Multimedia package base on 6.6.52 #2055
Conversation
| " | ||
| SRC_URI:append:imx-nxp-bsp = "${IMX_PATCH}" | ||
|
|
||
| PACKAGE_ARCH:imx-nxp-bsp = "${MACHINE_SOCARCH}" | ||
|
|
||
| GLIBC_64BIT_TIME_FLAGS = "" | ||
| GLIBC_64BIT_TIME_FLAGS:arm:imx-nxp-bsp = " -D_TIME_BITS=64 -D_FILE_OFFSET_BITS=64" |
There was a problem hiding this comment.
Is there any reason why we need to explicitly deal 64-bit time flags? Why we couldn't use the standard handling inside of the Yocto Project?
There was a problem hiding this comment.
This is a workaround to fix Y2038 issue, BSP shall build with time64 as line15 set with empty ,which means the timer flag not is 64bit. We can waiting for root fix and it shouldn't upstream now. Updated!
There was a problem hiding this comment.
But this is exactly what I'm concerning about, why we are disabling the build with 64-bit support. My question is if the correct fix wouldn't be remove the line 15.
|
I would like to clarify a comment I made earlier. It would be beneficial to improve the comments in each commit log. Specifically, I suggest describing the individual changes made to each recipe within their respective commit logs. This detailed documentation will be useful for referencing the reasons behind each change in the future. |
Override arm32 build time flag Signed-off-by: Zelan Zou <[email protected]>
Signed-off-by: Zelan Zou <[email protected]>
Signed-off-by: Zelan Zou <[email protected]>
Signed-off-by: Zelan Zou <[email protected]>
Bump version 2.1.8 -> 2.1.9 Signed-off-by: Zelan Zou <[email protected]>
Bump version 0.5.0 -> 0.6.0 Signed-off-by: Zelan Zou <[email protected]>
Bump version 4.9.1 -> 4.9.2 Signed-off-by: Zelan Zou <[email protected]>
Signed-off-by: Zelan Zou <[email protected]>
Drop md5sum check New feature: https://docs.yoctoproject.org/dev/singleindex.html#fetching-code https://docs.yoctoproject.org/dev/singleindex.html#src-uri-checksum-behaviour Signed-off-by: Zelan Zou <[email protected]>
zelan-nxp
force-pushed
the
multimedia-package-6.6.52
branch
from
876ea2c to
bf8e076
Compare
| require gstreamer1.0-plugins-common.inc | ||
| require gstreamer1.0-plugins-license.inc | ||
| require recipes-multimedia/gstreamer/gstreamer1.0-plugins-common.inc | ||
| require recipes-multimedia/gstreamer/gstreamer1.0-plugins-license.inc |
There was a problem hiding this comment.
This is half a change. We do this in meta-imx so we do not have to copy these .inc files locally. I suspect that @otavio will not want this, so probably this change should be removed.
There was a problem hiding this comment.
The idea of having those copied is to avoid upstream changes to change the behavior of the recipes that we are including here.
That said, we need to verify if there is any changes made upstream that we need to copy here as well.
| # it uses nested functions sadly, in ext/wayland/gstwaylandsink.c for GST_ELEMENT_REGISTER_DEFINE | ||
| # | ||
| TOOLCHAIN = "gcc" | ||
|
|
There was a problem hiding this comment.
Not sure whether this can be removed. It appears that we have been building without this change all along, which makes me think that we do not build whatever causes the problem. It might be better to not upstream this and instead to align meta-imx with this change.
There was a problem hiding this comment.
This is likely to avoid using other toolchain in case it is chosen as default. For example, it is likely to fix a problem if the distribution uses CLang as default toolchain.
There was a problem hiding this comment.
Ah, right, so bottom line this is backwards again. @zelan-nxp, please remove this from the PR and add it to meta-imx.
| @@ -5,7 +5,7 @@ | |||
| ########### OE-core copy ################## | |||
| # Upstream hash: 937817e5164f8af8452aec03ae3c45cb23d63df9 | |||
|
|
|||
| require gstreamer1.0-plugins-common.inc | |||
| require recipes-multimedia/gstreamer/gstreamer1.0-plugins-common.inc | |||
| @@ -5,7 +5,7 @@ | |||
| ########### OE-core copy ################## | |||
| # Upstream hash: 937817e5164f8af8452aec03ae3c45cb23d63df9 | |||
|
|
|||
| require gstreamer1.0-plugins-common.inc | |||
| require recipes-multimedia/gstreamer/gstreamer1.0-plugins-common.inc | |||
| file://0003-tests-use-a-dictionaries-for-environment.patch \ | ||
| file://0004-tests-add-helper-script-to-run-the-installed_tests.patch \ | ||
| file://0003-tests-use-a-dictionaries-for-environment.patch;striplevel=3 \ | ||
| file://0004-tests-add-helper-script-to-run-the-installed_tests.patch;striplevel=3 \ |
There was a problem hiding this comment.
This change was dropped upstream.
@zelan-nxp, please check that our recipes in meta-imx are aligned with the latest for the next release.
| @@ -60,7 +59,7 @@ do_package_qa[prefuncs] += "__set_insane_skip" | |||
|
|
|||
| python __split_libfslcodec_plugins() { | |||
| codecdir = bb.data.expand('${libdir}', d) | |||
| do_split_packages(d, codecdir, r'^lib_([^_]*).*_arm.*_elinux\.so\..*', | |||
| do_split_packages(d, codecdir, '^lib_([^_]*).*_arm.*_elinux\.so\..*', | |||
There was a problem hiding this comment.
This is the wrong way. The line in meta-freescale needs to be backported to meta-imx.
Upstream MM packages:
gstreamer: 1.24.0 ->1.24.7
dsp: 2.1.8 -> 2.1.9
opencl-converter: 0.5.0 ->0.6.0
imx-parser: 4.9.1- >4.9.2
new feature:
Drop md5sum check in recipe
https://docs.yoctoproject.org/dev/singleindex.html#fetching-code
https://docs.yoctoproject.org/dev/singleindex.html#src-uri-checksum-behaviour