devtool/upgrade: rename RECIPE_UPDATE_EXTRA_TASKS -> RECIPE_UPGRADE_EXTRA_TASKS
PATCH 1/11 devtool/upgrade: rename RECIPE_UPDATE_EXTRA_TASKS -> RECIPE_UPGRADE_EXTRA_TASKS
From: Alexander Kanavin <[email protected]>
'UPDATE' as a name is somewhat unfortunate as the variable is intended only for the 'devtool upgrade' operation and devtool also has an 'update-recipe' operation.
I also intend to introduce additional variables in RECIPE_UPGRADE_* space, so it's best to unify names now, while RECIPE_UPDATE_EXTRA_TASKS is very fresh.
--- meta/classes-recipe/cargo-update-recipe-crates.bbclass | 2 +- meta/lib/oeqa/selftest/cases/devtool.py | 2 +- scripts/lib/devtool/upgrade.py | 10 +++++----- 3 files changed, 7 insertions(+), 7 deletions(-)
meta/classes-recipe/cargo-update-recipe-crates.bbclass +1 -1
@@ -18,7 +18,7 @@ do_update_crates[depends] = "python3-native:do_populate_sysroot" do_update_crates[nostamp] = "1" do_update_crates[doc] = "Update the recipe by reading Cargo.lock and write in ${THISDIR}/${BPN}-crates.inc" -RECIPE_UPDATE_EXTRA_TASKS += "do_update_crates" +RECIPE_UPGRADE_EXTRA_TASKS += "do_update_crates" # The directory where to search for Cargo.lock files CARGO_LOCK_SRC_DIR ??= "${S}"
meta/lib/oeqa/selftest/cases/devtool.py +1 -1
@@ -2017,7 +2017,7 @@ class DevtoolUpgradeTests(DevtoolBase): newlines = f.readlines() self.assertEqual(desiredlines, newlines) - def test_devtool_upgrade_recipe_update_extra_tasks(self): + def test_devtool_upgrade_recipe_upgrade_extra_tasks(self): # Check preconditions self.assertTrue(not os.path.exists(self.workspacedir), 'This test cannot be run with a workspace directory under the build directory') self.track_for_cleanup(self.workspacedir)
scripts/lib/devtool/upgrade.py +5 -5
@@ -534,14 +534,14 @@ def _generate_license_diff(old_licenses, new_licenses): diff = diff + line return diff -def _run_recipe_update_extra_tasks(pn, rd, tinfoil): +def _run_recipe_upgrade_extra_tasks(pn, rd, tinfoil): tasks = [] - for task in (rd.getVar('RECIPE_UPDATE_EXTRA_TASKS') or '').split(): - logger.info('Running extra recipe update task: %s' % task) + for task in (rd.getVar('RECIPE_UPGRADE_EXTRA_TASKS') or '').split(): + logger.info('Running extra recipe upgrade task: %s' % task) res = tinfoil.build_targets(pn, task, handle_events=True) if not res: - raise DevtoolError('Running extra recipe update task %s for %s failed' % (task, pn)) + raise DevtoolError('Running extra recipe upgrade task %s for %s failed' % (task, pn)) def upgrade(args, config, basepath, workspace): """Entry point for the devtool 'upgrade' subcommand"""
@@ -618,7 +618,7 @@ def upgrade(args, config, basepath, workspace): copied, config.workspace_path, rd) standard._add_md5(config, pn, af) - _run_recipe_update_extra_tasks(pn, rd, tinfoil) + _run_recipe_upgrade_extra_tasks(pn, rd, tinfoil) update_unlockedsigs(basepath, workspace, args.fixed_setup, [pn])
Signature · 1 line
2.39.2
PATCH 2/11 devtool/upgrade: add RECIPE_UPGRADE_FINISH_EXTRA_TASKS
From: Alexander Kanavin <[email protected]>
This is modeled on RECIPE_UPGRADE_EXTRA_TASKS with the difference being that the tasks run in 'devtool finish', and only when finishing an upgrade operation. This allows recipes to define custom operations, such as renaming additional supplementary recipes (such as *-initial, *-native etc), or any other custom handling that devtool cannot know about.
The idea is to teach glib, cmake, mesa, systemd, libva, rt-tests, go, rust, etc etc about upgrading themselves properly with this, improving success rates in AUH.
--- scripts/lib/devtool/standard.py | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)
scripts/lib/devtool/standard.py +4 -1
@@ -2088,7 +2088,6 @@ def _get_layer(layername, d): logger.warning("Consider using path instead of base name to specify layer:\n\t\t%s" % '\n\t\t'.join(layer_paths)) return os.path.abspath(layer_paths[0]) - def finish(args, config, basepath, workspace): """Entry point for the devtool 'finish' subcommand""" import bb
@@ -2120,6 +2119,7 @@ def finish(args, config, basepath, workspace): destlayerdir = _get_layer(args.destination, tinfoil.config_data) recipefile = rd.getVar('FILE') + upgrade_finish_tasks = (rd.getVar('RECIPE_UPGRADE_FINISH_EXTRA_TASKS') or '').split() recipedir = os.path.dirname(recipefile) origlayerdir = oe.recipeutils.find_layerdir(recipefile)
@@ -2265,6 +2265,9 @@ def finish(args, config, basepath, workspace): else: _reset([args.recipename], no_clean=no_clean, remove_work=remove_work, config=config, basepath=basepath, workspace=workspace) + if removing_original: + for t in upgrade_finish_tasks: + exec_build_env_command(config.init_path, basepath, 'bitbake -c {} {}'.format(t, args.recipename)) return 0
Signature · 1 line
2.39.2
Re: [OE-core] [PATCH 02/11] devtool/upgrade: add RECIPE_UPGRADE_FINISH_EXTRA_TASKS
Quoted 9 lines · Alexander Kanavin via lists.openembedded.org
From: Alexander Kanavin <[email protected]>
This is modeled on
RECIPE_UPGRADE_EXTRA_TASKSwith the difference being that the tasks run in 'devtool finish', and only when finishing an upgrade operation. This allows recipes to define custom operations, such as renaming additional supplementary recipes (such as *-initial, *-native etc), or any other custom handling that devtool cannot know about.
Quoted 3 lines
The idea is to teach glib, cmake, mesa, systemd, libva, rt-tests, go, rust, etc etc about upgrading themselves properly with this, improving success rates in AUH.
Whilst I understand the desire, I'm not sure this lives up to way devtool works.
By that I mean that devtool upgrade usually does testing of the result but with what you implement here, the second recipe is never actually tested so things could easily break after devtool finish and you wouldn't know without further testing.
Thinking about this a bit more, we should be able to ask bitbake which other recipes "share" files since I think the list of included files is actually cached by bitbake. Yes, we may need API but in theory this means devtool can know which other recipes it needs to upgrade at the same time, and when doing so, run them through the same process as normal upgrades.
So yes, it will be a bit more work but the end result should be more generic and scalable?
Cheers,
Richard
Re: [OE-core] [PATCH 02/11] devtool/upgrade: add RECIPE_UPGRADE_FINISH_EXTRA_TASKS
Quoted 7 lines · Richard Purdie
Whilst I understand the desire, I'm not sure this lives up to way devtool works.
By that I mean that devtool upgrade usually does testing of the result but with what you implement here, the second recipe is never actually tested so things could easily break after devtool finish and you wouldn't know without further testing.
Yes. Supplementary .bb renaming only works if those additional recipes do not add licensing statements, or custom patches (and AUH also won't test-build them separately). It was something of a low hanging fruit I wanted to try, as for things like mesa, qemu, cmake and others in the patchset I sent this does turn reliably failing AUH updates into working ones (as long as only the renaming is needed).
Quoted 9 lines
Thinking about this a bit more, we should be able to ask bitbake which other recipes "share" files since I think the list of included files is actually cached by bitbake. Yes, we may need API but in theory this means devtool can know which other recipes it needs to upgrade at the same time, and when doing so, run them through the same process as normal upgrades.
So yes, it will be a bit more work but the end result should be more generic and scalable?
I would first try to simply list the additional recipes directly in the main one:RECIPE_UPGRADE_ADDITIONAL = "qemu-native qemu-system-native" and see if devtool can be easily made to iterate over them:
devtool upgrade qemu
- internally performs upgrade qemu-native qemu-system-native
devtool finish qemu
- does the same for finish operation.
And since these share includes and patches and other items in SRC_URI, we need to ensure they don't step on each other. Although there will be multiple overwriting in 'devtool finish' which is okay as long as one updated thing is overwritten with the same updated thing.
Alex
Re: [OE-core] [PATCH 02/11] devtool/upgrade: add RECIPE_UPGRADE_FINISH_EXTRA_TASKS
Quoted 35 lines · Alexander Kanavin
Quoted 7 lines · Richard Purdie
Whilst I understand the desire, I'm not sure this lives up to way devtool works.
By that I mean that devtool upgrade usually does testing of the result but with what you implement here, the second recipe is never actually tested so things could easily break after devtool finish and you wouldn't know without further testing.
Yes. Supplementary .bb renaming only works if those additional recipes do not add licensing statements, or custom patches (and AUH also won't test-build them separately). It was something of a low hanging fruit I wanted to try, as for things like mesa, qemu, cmake and others in the patchset I sent this does turn reliably failing AUH updates into working ones (as long as only the renaming is needed).
Quoted 9 lines
Thinking about this a bit more, we should be able to ask bitbake which other recipes "share" files since I think the list of included files is actually cached by bitbake. Yes, we may need API but in theory this means devtool can know which other recipes it needs to upgrade at the same time, and when doing so, run them through the same process as normal upgrades.
So yes, it will be a bit more work but the end result should be more generic and scalable?
I would first try to simply list the additional recipes directly in the main one:
RECIPE_UPGRADE_ADDITIONAL= "qemu-native qemu-system-native" and see if devtool can be easily made to iterate over them:devtool upgrade qemu
- internally performs upgrade qemu-native qemu-system-native
Quoted 7 lines
devtool finish qemu
- does the same for finish operation.And since these share includes and patches and other items in
SRC_URI, we need to ensure they don't step on each other. Although there will be multiple overwriting in 'devtool finish' which is okay as long as one updated thing is overwritten with the same updated thing.
I don't think we even need to list the recipes like this. When devtool upgrade runs, we get a list of files which change. We can then query bitbake with that list of changed files and ask "which recipes also depend on these?".
Based upon that list we can then know what else might need to change (e.g. if any of those recipe names have the same PV in) and also which recipe targets also need to be tested as part of the upgrade.
Cheers,
Richard
Re: [OE-core] [PATCH 02/11] devtool/upgrade: add RECIPE_UPGRADE_FINISH_EXTRA_TASKS
Quoted 4 lines · Richard Purdie
I don't think we even need to list the recipes like this. When devtool upgrade runs, we get a list of files which change. We can then query bitbake with that list of changed files and ask "which recipes also depend on these?".
Right. Can you give me a lead on how this can be obtained? (e.g. where is the data structure it's in, or a function that returns it)
Alex
Re: [OE-core] [PATCH 02/11] devtool/upgrade: add RECIPE_UPGRADE_FINISH_EXTRA_TASKS
Quoted 10 lines · Alexander Kanavin
Quoted 6 lines · Richard Purdie
I don't think we even need to list the recipes like this. When devtool
upgrade runs, we get a list of files which change. We can then query
bitbake with that list of changed files and ask "which recipes also depend on these?".
Quoted 3 lines
Right. Can you give me a lead on how this can be obtained? (e.g. where
is the data structure it's in, or a function that returns it)
In tinfoil you can see the code in TinfoilCookerAdapter/TinfoilRecipeCacheAdapter i.e.:
self.recipecaches[mc].inherits
i.e. the result of:
attrvalue = self.tinfoil.run_command('getRecipeInherits', self.mc) or {}
I appreciate it calls this "inherits" but this is actually a list of any file included by the base recipe so any bbclass, .bb, or .inc file that was included should be listed here. The True/False associated with it is whether the file existed or not so that if it changed, we know to reparse.
Cheers,
Richard
Re: [OE-core] [PATCH 02/11] devtool/upgrade: add RECIPE_UPGRADE_FINISH_EXTRA_TASKS
Quoted 5 lines · Richard Purdie
I appreciate it calls this "inherits" but this is actually a list of any file included by the base recipe so any bbclass, .bb, or .inc file that was included should be listed here. The True/False associated with it is whether the file existed or not so that if it changed, we know to reparse.
This doesn't seem to be correct. I added
+ logger.info(tinfoil.cooker_data.inherits['/srv/work/alex/poky/meta/recipes-devtools/cmake/cmake_3.29.3.bb'])
+ logger.info(tinfoil.cooker_data.inherits['/srv/work/alex/poky/meta/recipes-devtools/cmake/cmake-native_3.29.3.bb'])
to devtool to see how the data in that attribute looks like, and unfortunately it contains only .bbclass filenames and not .inc or any other dependent files:
INFO: ['/srv/work/alex/poky/meta/classes-global/base.bbclass', '/srv/work/alex/poky/meta/classes-global/patch.bbclass', '/srv/work/alex/poky/meta/classes/terminal.bbclass', '/srv/work/alex/poky/meta/classes-global/staging.bbclass', '/srv/work/alex/poky/meta/classes-global/mirrors.bbclass', '/srv/work/alex/poky/meta/classes-global/utils.bbclass', '/srv/work/alex/poky/meta/classes-global/utility-tasks.bbclass', ...
etc. - no cmake.inc mentioned anywhere.
So assuming I didn't miss something, this either needs further bitbake/tinfoil extensions (even more work), or we can just write manual hints for lockstep upgrades into recipes (there's not that many of them so I'm leaning that way tbh).
Alex
Re: [OE-core] [PATCH 02/11] devtool/upgrade: add RECIPE_UPGRADE_FINISH_EXTRA_TASKS
Quoted 34 lines · Alexander Kanavin
Quoted 9 lines · Richard Purdie
I appreciate it calls this "inherits" but this is actually a list of
any file included by the base recipe so any bbclass, .bb, or .inc file
that was included should be listed here. The True/False associated with
it is whether the file existed or not so that if it changed, we know to
reparse.This doesn't seem to be correct. I added
+
logger.info(tinfoil.cooker_data.inherits['/srv/work/alex/poky/meta/re cipes-devtools/cmake/cmake_3.29.3.bb'])
+ logger.info(tinfoil.cooker_data.inherits['/srv/work/alex/poky/meta/re cipes-devtools/cmake/cmake-native_3.29.3.bb'])to devtool to see how the data in that attribute looks like, and unfortunately it contains only .bbclass filenames and not .inc or any other dependent files:
INFO: ['
/srv/work/alex/poky/meta/classes-global/base.bbclass', '/srv/work/alex/poky/meta/classes-global/patch.bbclass', '/srv/work/alex/poky/meta/classes/terminal.bbclass', '/srv/work/alex/poky/meta/classes-global/staging.bbclass', '/srv/work/alex/poky/meta/classes-global/mirrors.bbclass', '/srv/work/alex/poky/meta/classes-global/utils.bbclass', '/srv/work/alex/poky/meta/classes-global/utility-tasks.bbclass', ...
Quoted 7 lines
etc. - no cmake.inc mentioned anywhere.
So assuming I didn't miss something, this either needs further bitbake/tinfoil extensions (even more work), or we can just write manual hints for lockstep upgrades into recipes (there's not that many
of them so I'm leaning that way tbh).
Sorry, I've got confused. Look at tinfoil.cooker_data.file_depends
Cheers,
Richard
Re: [OE-core] [PATCH 02/11] devtool/upgrade: add RECIPE_UPGRADE_FINISH_EXTRA_TASKS
Quoted 1 line · Richard Purdie
Sorry, I've got confused. Look at tinfoil.cooker_data.file_depends
AttributeError: TinfoilRecipeCacheAdapter instance has no attribute 'file_depends'
I had looked into the code, it seems that data is internal to bitbake, and isn't exposed to tinfoil.
It's also placed into BBINCLUDED per-recipe, but that would require iterating over "recipes = tinfoil.all_recipe_files(variants=False)" in every upgrade operation, which might slow things down a bit too much. Worth trying though?
Alex
Re: [OE-core] [PATCH 02/11] devtool/upgrade: add RECIPE_UPGRADE_FINISH_EXTRA_TASKS
Quoted 10 lines · Alexander Kanavin
Quoted 1 line · Richard Purdie
Sorry, I've got confused. Look at tinfoil.cooker_data.file_depends
AttributeError: TinfoilRecipeCacheAdapter instance has no attribute '
file_depends'I had looked into the code, it seems that data is internal to bitbake, and isn't exposed to tinfoil.
We might consider exposing it?
There data is there and in the cache so in theory we should be able to do something with it, even if we filter the data so we don't expose the internal data structure (only list files that exist via tinfoil).
Cheers,
Richard
PATCH 3/11 recipe-upgrade-additional-rename.bbclass: add a class
In reply to Alexander KanavinFrom: Alexander Kanavin <[email protected]>
Please see the comment in the class for the rationale.
--- .../recipe-upgrade-additional-rename.bbclass | 22 +++++++++++++++++++ 1 file changed, 22 insertions(+) create mode 100644 meta/classes-recipe/recipe-upgrade-additional-rename.bbclass
meta/classes-recipe/recipe-upgrade-additional-rename.bbclass new file +22 -0
@@ -0,0 +1,22 @@ +# This class can be used to add file renames in devtool version upgrades +# operations when there are supplemental recipes that need to be updated +# in lockstep with the main one by renaming the files. This is common, +# e.g. glib, cmake, mesa, libva, etc have this issue. +# +# Example (from glib which has glib-initial to handle): +# +# inherit recipe-upgrade-additional-rename +# RECIPE_UPGRADE_ADDITIONAL_RENAME = "glib-2.0-initial_{pv}.bb" +# +RECIPE_UPGRADE_FINISH_EXTRA_TASKS += "do_recipe_upgrade_additional_rename" +addtask do_recipe_upgrade_additional_rename + +python do_recipe_upgrade_additional_rename() { + import glob + origpath = d.getVar("THISDIR") + newpv = d.getVar("PV") + + for f_pattern in d.getVar("RECIPE_UPGRADE_ADDITIONAL_RENAME").split(): + for f_disk in glob.glob(os.path.join(origpath, f_pattern.format(pv="*"))): + os.rename(f_disk, os.path.join(origpath, f_pattern.format(pv=newpv))) +}
Signature · 1 line
2.39.2
PATCH 4/11 glib-2.0: rename glib-initial recipe as well on version upgrades
In reply to Alexander KanavinFrom: Alexander Kanavin <[email protected]>
This makes use of all the facilities introduced in the previous commits.
--- meta/recipes-core/glib-2.0/glib-2.0-initial_2.80.2.bb | 2 +- meta/recipes-core/glib-2.0/glib-2.0_2.80.2.bb | 2 ++ 2 files changed, 3 insertions(+), 1 deletion(-)
meta/recipes-core/glib-2.0/glib-2.0-initial_2.80.2.bb +1 -1
@@ -2,4 +2,4 @@ require glib.inc PACKAGES = "" PACKAGECONFIG = "" - +RECIPE_NO_UPDATE_REASON = "Upgrades are handled in base glib-2.0 recipe"
meta/recipes-core/glib-2.0/glib-2.0_2.80.2.bb +2 -0
@@ -1,2 +1,4 @@ require glib.inc +inherit recipe-upgrade-additional-rename +RECIPE_UPGRADE_ADDITIONAL_RENAME = "glib-2.0-initial_{pv}.bb"
Signature · 1 line
2.39.2
PATCH 5/11 cmake: rename cmake-native recipe on version upgrades
In reply to Alexander KanavinFrom: Alexander Kanavin <[email protected]>
--- meta/recipes-devtools/cmake/cmake_3.29.3.bb | 3 +++ 1 file changed, 3 insertions(+)
meta/recipes-devtools/cmake/cmake_3.29.3.bb +3 -0
@@ -66,3 +66,6 @@ FILES:${PN}-doc += "${docdir}/cmake-${CMAKE_MAJOR_VERSION}" FILES:${PN}-dev = "" BBCLASSEXTEND = "nativesdk" + +inherit recipe-upgrade-additional-rename +RECIPE_UPGRADE_ADDITIONAL_RENAME = "cmake-native_{pv}.bb"
Signature · 1 line
2.39.2
PATCH 6/11 qemu: rename qemu-native and qemu-system-native recipes on version upgrades
In reply to Alexander KanavinFrom: Alexander Kanavin <[email protected]>
--- meta/recipes-devtools/qemu/qemu_9.0.0.bb | 3 +++ 1 file changed, 3 insertions(+)
meta/recipes-devtools/qemu/qemu_9.0.0.bb +3 -0
@@ -25,3 +25,6 @@ PACKAGECONFIG:class-nativesdk ??= "fdt sdl kvm pie slirp \ " # ppc32 hosts are no longer supported in qemu COMPATIBLE_HOST:powerpc = "null" + +inherit recipe-upgrade-additional-rename +RECIPE_UPGRADE_ADDITIONAL_RENAME = "qemu-native_{pv}.bb qemu-system-native_{pv}.bb"
Signature · 1 line
2.39.2
PATCH 7/11 libva: rename libva-initial recipe on recipe version upgrades
In reply to Alexander KanavinFrom: Alexander Kanavin <[email protected]>
--- meta/recipes-graphics/libva/libva-initial_2.20.0.bb | 1 + meta/recipes-graphics/libva/libva_2.20.0.bb | 3 +++ 2 files changed, 4 insertions(+)
meta/recipes-graphics/libva/libva-initial_2.20.0.bb +1 -0
@@ -6,3 +6,4 @@ do_install:append () { rm -f ${D}${libdir}/*.so* } +RECIPE_NO_UPDATE_REASON = "Upgrades are handled in base libva recipe"
meta/recipes-graphics/libva/libva_2.20.0.bb +3 -0
@@ -13,3 +13,6 @@ RDEPENDS:${PN}-glx =+ "${PN}-x11" FILES:${PN}-x11 =+ "${libdir}/libva-x11*${SOLIBS}" FILES:${PN}-glx =+ "${libdir}/libva-glx*${SOLIBS}" FILES:${PN}-wayland =+ "${libdir}/libva-wayland*${SOLIBS}" + +inherit recipe-upgrade-additional-rename +RECIPE_UPGRADE_ADDITIONAL_RENAME = "libva-initial_{pv}.bb"
Signature · 1 line
2.39.2
PATCH 8/11 util-linux: rename util-linux-libuuid on recipe version upgrades
In reply to Alexander KanavinFrom: Alexander Kanavin <[email protected]>
--- meta/recipes-core/util-linux/util-linux-libuuid_2.39.3.bb | 2 ++ meta/recipes-core/util-linux/util-linux_2.39.3.bb | 3 +++ 2 files changed, 5 insertions(+)
meta/recipes-core/util-linux/util-linux-libuuid_2.39.3.bb +2 -0
@@ -17,3 +17,5 @@ do_install:append() { } BBCLASSEXTEND = "native nativesdk" + +RECIPE_NO_UPDATE_REASON = "Upgrades are handled in base util-linux recipe"
meta/recipes-core/util-linux/util-linux_2.39.3.bb +3 -0
@@ -346,3 +346,6 @@ do_install_ptest:append:libc-musl() { rm -rf ${D}${PTEST_PATH}/$t done } + +inherit recipe-upgrade-additional-rename +RECIPE_UPGRADE_ADDITIONAL_RENAME = "util-linux-libuuid_{pv}.bb"
Signature · 1 line
2.39.2
PATCH 9/11 systemd: rename systemd-boot on recipe version upgrades
In reply to Alexander KanavinFrom: Alexander Kanavin <[email protected]>
--- meta/recipes-core/systemd/systemd-boot_255.6.bb | 1 + meta/recipes-core/systemd/systemd_255.6.bb | 3 +++ 2 files changed, 4 insertions(+)
meta/recipes-core/systemd/systemd-boot_255.6.bb +1 -0
@@ -65,3 +65,4 @@ do_deploy () { addtask deploy before do_build after do_compile +RECIPE_NO_UPDATE_REASON = "Upgrades are handled in base systemd recipe"
meta/recipes-core/systemd/systemd_255.6.bb +3 -0
@@ -899,3 +899,6 @@ pkg_postinst:udev-hwdb () { pkg_prerm:udev-hwdb () { rm -f $D${sysconfdir}/udev/hwdb.bin } + +inherit recipe-upgrade-additional-rename +RECIPE_UPGRADE_ADDITIONAL_RENAME = "systemd-boot_{pv}.bb systemd-boot-native_{pv}.bb"
Signature · 1 line
2.39.2
PATCH 10/11 mesa: rename mesa-gl on recipe version upgrades
In reply to Alexander KanavinFrom: Alexander Kanavin <[email protected]>
--- meta/recipes-graphics/mesa/mesa_24.0.7.bb | 2 ++ 1 file changed, 2 insertions(+)
meta/recipes-graphics/mesa/mesa_24.0.7.bb +2 -0
@@ -1,2 +1,4 @@ require ${BPN}.inc +inherit recipe-upgrade-additional-rename +RECIPE_UPGRADE_ADDITIONAL_RENAME = "mesa-gl_{pv}.bb"
Signature · 1 line
2.39.2
PATCH 11/11 rt-tests: rename hwlatdetect on recipe version upgrades
In reply to Alexander KanavinFrom: Alexander Kanavin <[email protected]>
--- meta/recipes-rt/rt-tests/hwlatdetect_git.bb | 2 ++ meta/recipes-rt/rt-tests/rt-tests_git.bb | 3 +++ 2 files changed, 5 insertions(+)
meta/recipes-rt/rt-tests/hwlatdetect_git.bb +2 -0
@@ -24,3 +24,5 @@ do_install() { FILES:${PN} += "${libdir}/python${PYTHON_BASEVERSION}/dist-packages/hwlatdetect.py" RDEPENDS:${PN} = "python3-core " RRECOMMENDS:${PN} = "kernel-module-hwlat-detector" + +RECIPE_NO_UPDATE_REASON = "Upgrades are handled in rt-tests recipe"
meta/recipes-rt/rt-tests/rt-tests_git.bb +3 -0
@@ -34,3 +34,6 @@ RDEPENDS:${PN}-ptest += " stress-ng python3 python3-multiprocessing python3-date FILES:${PN} += "${prefix}/src/backfire" RDEPENDS:${PN} += "bash" + +inherit recipe-upgrade-additional-rename +RECIPE_UPGRADE_ADDITIONAL_RENAME = "hwlatdetect_{pv}.bb"
Signature · 1 line
2.39.2