adwaita-icon-theme: 43 -> 45.0
PATCH v2 1/1 adwaita-icon-theme: 43 -> 45.0
From: Kai Kang <[email protected]>
Update adwaita-icon-theme from 43 to 45.0.
* meson is used to replace autotools, then the patch is redundant
* not overwrite PACKAGES and split license files to ${PN}-doc
---
v2:
* not overwrite PACKAGES and split license files to ${PN}-doc
.../0001-Don-t-use-AC_CANONICAL_HOST.patch | 27 -------------------
...theme_43.bb => adwaita-icon-theme_45.0.bb} | 11 ++++----
2 files changed, 6 insertions(+), 32 deletions(-)
delete mode 100644 meta/recipes-gnome/gnome/adwaita-icon-theme/0001-Don-t-use-AC_CANONICAL_HOST.patch
rename meta/recipes-gnome/gnome/{adwaita-icon-theme_43.bb => adwaita-icon-theme_45.0.bb} (84%)meta/recipes-gnome/gnome/adwaita-icon-theme/0001-Don-t-use-AC_CANONICAL_HOST.patch deleted +0 -27
@@ -1,27 +0,0 @@ -From e1a12b73c25d1ad0f267b22ac5d799bc1fbd0fa4 Mon Sep 17 00:00:00 2001 -From: Jussi Kukkonen <[email protected]> -Date: Tue, 30 May 2017 14:55:49 +0300 -Subject: [PATCH] Don't use AC_CANONICAL_HOST - -This won't work when building allarch (and is only used to find out if -target is windows). - -Upstream-Status: Inappropriate [embedded specific] -Signed-off-by: Jussi Kukkonen <[email protected]> - ---- - configure.ac | 1 - - 1 file changed, 1 deletion(-) - -diff --git a/configure.ac b/configure.ac -index 6ddc49b..39752bb 100644 ---- a/configure.ac -+++ b/configure.ac -@@ -3,7 +3,6 @@ AC_PREREQ(2.53) - - AC_INIT([adwaita-icon-theme], [43], - [http://bugzilla.gnome.org/enter_bug.cgi?product=adwaita-icon-theme]) --AC_CANONICAL_HOST - AC_CONFIG_MACRO_DIR([m4]) - AC_CONFIG_SRCDIR([index.theme.in]) -
meta/recipes-gnome/gnome/adwaita-icon-theme_45.0.bb renamed from meta/recipes-gnome/gnome/adwaita-icon-theme_43.bb +6 -5
@@ -10,15 +10,14 @@ LIC_FILES_CHKSUM = "file://COPYING;md5=c84cac88e46fc07647ea07e6c24eeb7c \ file://COPYING_CCBYSA3;md5=96143d33de3a79321b1006c4e8ed07e7 \ file://COPYING_LGPL;md5=e6a600fd5e1d9cbde2d983680233ad02" -inherit allarch autotools pkgconfig gettext gtk-icon-cache gnomebase +GNOMEBASEBUILDCLASS = "meson" +inherit allarch gnomebase gtk-icon-cache -SRC_URI += " file://0001-Don-t-use-AC_CANONICAL_HOST.patch" - -SRC_URI[archive.sha256sum] = "2e3ac77d32a6aa5554155df37e8f0a0dd54fc5a65fd721e88d505f970da32ec6" +SRC_URI[archive.sha256sum] = "2442bfb06f4e6cc95bf6e2682fdff98fa5eddc688751b9d6215c623cb4e42ff1" DEPENDS += "librsvg-native" -PACKAGES = "${PN}-cursors ${PN}-symbolic-hires ${PN}-symbolic ${PN}-hires ${PN}" +PACKAGES =+ "${PN}-cursors ${PN}-symbolic-hires ${PN}-symbolic ${PN}-hires" RREPLACES:${PN} = "gnome-icon-theme" RCONFLICTS:${PN} = "gnome-icon-theme"
@@ -37,4 +36,6 @@ FILES:${PN}-hires = "${prefix}/share/icons/Adwaita/256x256/ \ FILES:${PN} = "${prefix}/share/icons/Adwaita/ \ ${prefix}/share/pkgconfig/adwaita-icon-theme.pc" +FILES:${PN}-doc = "${prefix}/share/licenses/adwaita-icon-theme" + BBCLASSEXTEND = "native nativesdk"
Signature · 1 line
2.34.1
Did you verify that fail in
https://autobuilder.yoctoproject.org/typhoon/#/builders/79/builds/5795/steps/14/logs/stdio
does not occur?
Also booting sato and checking that it look ok is good to do.
Alex
Quoted 89 lines · On Thu, 21 Sept 2023 at 08:47, <[email protected]> wrote:
From: Kai Kang <[email protected]>
Update adwaita-icon-theme from 43 to 45.0.
* meson is used to replace autotools, then the patch is redundant
* not overwrite PACKAGES and split license files to ${PN}-docSigned-off-by Kai Kang <[email protected]>---
v2:
* not overwrite PACKAGES and split license files to ${PN}-doc.../0001-Don-t-use-AC_CANONICAL_HOST.patch | 27 ------------------- ...theme_43.bb => adwaita-icon-theme_45.0.bb} | 11 ++++---- 2 files changed, 6 insertions(+), 32 deletions(-) delete mode 100644 meta/recipes-gnome/gnome/adwaita-icon-theme/0001-Don-t-use-AC_CANONICAL_HOST.patch rename meta/recipes-gnome/gnome/{adwaita-icon-theme_43.bb => adwaita-icon-theme_45.0.bb} (84%)meta/recipes-gnome/gnome/adwaita-icon-theme/0001-Don-t-use-AC_CANONICAL_HOST.patch deleted +0
-27@@ -1,27 +0,0 @@ -From e1a12b73c25d1ad0f267b22ac5d799bc1fbd0fa4 Mon Sep 17 00:00:00 2001 -From: Jussi Kukkonen <[email protected]> -Date: Tue, 30 May 2017 14:55:49 +0300 -Subject: [PATCH] Don't use AC_CANONICAL_HOST - -This won't work when building allarch (and is only used to find out if -target is windows). - -Upstream-Status: Inappropriate [embedded specific] -Signed-off-by: Jussi Kukkonen <[email protected]> - ---- - configure.ac | 1 - - 1 file changed, 1 deletion(-) - -diff --git a/configure.ac b/configure.ac -index 6ddc49b..39752bb 100644 ---- a/configure.ac -+++ b/configure.ac -@@ -3,7 +3,6 @@ AC_PREREQ(2.53) - - AC_INIT([adwaita-icon-theme], [43], - [http://bugzilla.gnome.org/enter_bug.cgi?product=adwaita-icon-theme]) --AC_CANONICAL_HOST - AC_CONFIG_MACRO_DIR([m4]) - AC_CONFIG_SRCDIR([index.theme.in]) -meta/recipes-gnome/gnome/adwaita-icon-theme_45.0.bb renamed from meta/recipes-gnome/gnome/adwaita-icon-theme_43.bb +6
-6@@ -10,15 +10,14 @@ LIC_FILES_CHKSUM = "file://COPYING;md5=c84cac88e46fc07647ea07e6c24eeb7c \ file://COPYING_CCBYSA3;md5=96143d33de3a79321b1006c4e8ed07e7 \ file://COPYING_LGPL;md5=e6a600fd5e1d9cbde2d983680233ad02" -inherit allarch autotools pkgconfig gettext gtk-icon-cache gnomebase +GNOMEBASEBUILDCLASS = "meson" +inherit allarch gnomebase gtk-icon-cache -SRC_URI += " file://0001-Don-t-use-AC_CANONICAL_HOST.patch" - -SRC_URI[archive.sha256sum] = "2e3ac77d32a6aa5554155df37e8f0a0dd54fc5a65fd721e88d505f970da32ec6" +SRC_URI[archive.sha256sum] = "2442bfb06f4e6cc95bf6e2682fdff98fa5eddc688751b9d6215c623cb4e42ff1" DEPENDS += "librsvg-native" -PACKAGES = "${PN}-cursors ${PN}-symbolic-hires ${PN}-symbolic ${PN}-hires ${PN}" +PACKAGES =+ "${PN}-cursors ${PN}-symbolic-hires ${PN}-symbolic ${PN}-hires" RREPLACES:${PN} = "gnome-icon-theme" RCONFLICTS:${PN} = "gnome-icon-theme"@@ -37,4 +36,6 @@ FILES:${PN}-hires = "${prefix}/share/icons/Adwaita/256x256/ \ FILES:${PN} = "${prefix}/share/icons/Adwaita/ \ ${prefix}/share/pkgconfig/adwaita-icon-theme.pc" +FILES:${PN}-doc = "${prefix}/share/licenses/adwaita-icon-theme" + BBCLASSEXTEND = "native nativesdk" --2.34.1
Quoted 3 lines · Alexander Kanavin
Did you verify that fail in
https://autobuilder.yoctoproject.org/typhoon/#/builders/79/builds/5795/steps/14/logs/stdio
does not occur?
Yes, it is caused by the update. I'll check the root cause.
Regards,
Kai
Quoted 94 lines
Also booting sato and checking that it look ok is good to do.
Alex
Quoted 88 lines · On Thu, 21 Sept 2023 at 08:47, <[email protected]> wrote:
From: Kai Kang <[email protected]>
Update adwaita-icon-theme from 43 to 45.0.
* meson is used to replace autotools, then the patch is redundant
* not overwrite PACKAGES and split license files to ${PN}-docSigned-off-by Kai Kang <[email protected]>---
v2:
* not overwrite PACKAGES and split license files to ${PN}-doc.../0001-Don-t-use-AC_CANONICAL_HOST.patch | 27 ------------------- ...theme_43.bb => adwaita-icon-theme_45.0.bb} | 11 ++++---- 2 files changed, 6 insertions(+), 32 deletions(-)delete mode 100644
meta/recipes-gnome/gnome/adwaita-icon-theme/0001-Don-t-use-AC_CANONICAL_HOST.patchrename meta/recipes-gnome/gnome/{adwaita-icon-theme_43.bb => adwaita-icon-theme_45.0.bb} (84%)meta/recipes-gnome/gnome/adwaita-icon-theme/0001-Don-t-use-AC_CANONICAL_HOST.patch deleted +0
-27@@ -1,27 +0,0 @@ -From e1a12b73c25d1ad0f267b22ac5d799bc1fbd0fa4 Mon Sep 17 00:00:00 2001 -From: Jussi Kukkonen <[email protected]> -Date: Tue, 30 May 2017 14:55:49 +0300 -Subject: [PATCH] Don't use AC_CANONICAL_HOST - -This won't work when building allarch (and is only used to find out if -target is windows). - -Upstream-Status: Inappropriate [embedded specific] -Signed-off-by: Jussi Kukkonen <[email protected]> - ---- - configure.ac | 1 - - 1 file changed, 1 deletion(-) - -diff --git a/configure.ac b/configure.ac -index 6ddc49b..39752bb 100644 ---- a/configure.ac -+++ b/configure.ac -@@ -3,7 +3,6 @@ AC_PREREQ(2.53) - - AC_INIT([adwaita-icon-theme], [43], - [http://bugzilla.gnome.org/enter_bug.cgi?product=adwaita-icon-theme]) --AC_CANONICAL_HOST - AC_CONFIG_MACRO_DIR([m4]) - AC_CONFIG_SRCDIR([index.theme.in]) -meta/recipes-gnome/gnome/adwaita-icon-theme_45.0.bb renamed from meta/recipes-gnome/gnome/adwaita-icon-theme_43.bb +6
-6@@ -10,15 +10,14 @@ LIC_FILES_CHKSUM = "file://COPYING;md5=c84cac88e46fc07647ea07e6c24eeb7c \ file://COPYING_CCBYSA3;md5=96143d33de3a79321b1006c4e8ed07e7 \ file://COPYING_LGPL;md5=e6a600fd5e1d9cbde2d983680233ad02" -inherit allarch autotools pkgconfig gettext gtk-icon-cache gnomebase +GNOMEBASEBUILDCLASS = "meson" +inherit allarch gnomebase gtk-icon-cache -SRC_URI += " file://0001-Don-t-use-AC_CANONICAL_HOST.patch" - -SRC_URI[archive.sha256sum] = "2e3ac77d32a6aa5554155df37e8f0a0dd54fc5a65fd721e88d505f970da32ec6" +SRC_URI[archive.sha256sum] = "2442bfb06f4e6cc95bf6e2682fdff98fa5eddc688751b9d6215c623cb4e42ff1" DEPENDS += "librsvg-native" -PACKAGES = "${PN}-cursors ${PN}-symbolic-hires ${PN}-symbolic ${PN}-hires ${PN}" +PACKAGES =+ "${PN}-cursors ${PN}-symbolic-hires ${PN}-symbolic ${PN}-hires" RREPLACES:${PN} = "gnome-icon-theme" RCONFLICTS:${PN} = "gnome-icon-theme"@@ -37,4 +36,6 @@ FILES:${PN}-hires = "${prefix}/share/icons/Adwaita/256x256/ \ FILES:${PN} = "${prefix}/share/icons/Adwaita/ \ ${prefix}/share/pkgconfig/adwaita-icon-theme.pc" +FILES:${PN}-doc = "${prefix}/share/licenses/adwaita-icon-theme" + BBCLASSEXTEND = "native nativesdk" --2.34.1
Signature · 2 lines
Kai Kang Wind River Linux
Quoted 1 line · Kai Kang
Yes, it is caused by the update. I'll check the root cause.
Maybe it is somehow related to ${prefix}/share and replacing by ${datadir} would help?
This fixed an issue I had with gnome-themes-extra
<https://git.openembedded.org/meta-openembedded/commit/?h=master-next&id=ab1f6ee8bf4835fabdf02cb0e30b20e4ff11e5fb>
Quoted 4 lines · Markus Volk
Maybe it is somehow related to ${prefix}/share and replacing by ${datadir} would help?
This fixed an issue I had with gnome-themes-extra
https://git.openembedded.org/meta-openembedded/commit/?h=master-next&id=ab1f6ee8bf4835fabdf02cb0e30b20e4ff11e5fb
Things diverge much earlier:
- 'adwaita-icon-theme/do_write_config':
'cdb1f880efd08a498798793a1cbc51e48272a71b2ec48b71f263b7af9c8b8ca9',
+ 'adwaita-icon-theme/do_write_config':
'690686d4d2721c21520f4736e3718e414f70366bea48b2df805b37a6a5193a1b',
Something in that task is contaminated by target-specific value.
Alex
GTKIC_VERSION = "4"
this line might be added to use gtk4 update-gtk-icon-cache ?
and probably also the DEPENDS line could be removed ?
Quoted 17 lines · Alexander Kanavin
Quoted 6 lines · On Thu, 21 Sept 2023 at 10:34, Markus Volk <[email protected]…
Maybe it is somehow related to ${prefix}/share and replacing by
${datadir} would help?This fixed an issue I had with gnome-themes-extra
Things diverge much earlier:
- 'adwaita-icon-theme/do_write_config':
'cdb1f880efd08a498798793a1cbc51e48272a71b2ec48b71f263b7af9c8b8ca9',
+ 'adwaita-icon-theme/do_write_config':
'690686d4d2721c21520f4736e3718e414f70366bea48b2df805b37a6a5193a1b',Something in that task is contaminated by target-specific value.
Alex
Markus, please. This is shotgun debugging. If you'd like to help, it would be more productive to take the patch and build it locally to have first hand evidence of what is happening. The failing test doesn't actually require a heavy rebuild, as it works only from task signatures.
Alex
Quoted 13 lines · Markus Volk
GTKIC_VERSION= "4"
this line might be added to use gtk4 update-gtk-icon-cache ?and probably also the DEPENDS line could be removed ?
On Thu, Sep 21 2023 at 10:40:49 AM +02:00:00, Alexander Kanavin <[email protected]> wrote:
On Thu, 21 Sept 2023 at 10:34, Markus Volk <[email protected]> wrote:
Maybe it is somehow related to ${prefix}/share and replacing by ${datadir} would help? This fixed an issue I had with gnome-themes-extra https://git.openembedded.org/meta-openembedded/commit/?h=master-next&id=ab1f6ee8bf4835fabdf02cb0e30b20e4ff11e5fb
Things diverge much earlier: - 'adwaita-icon-theme/do_write_config': '
cdb1f880efd08a498798793a1cbc51e48272a71b2ec48b71f263b7af9c8b8ca9', + 'adwaita-icon-theme/do_write_config': '690686d4d2721c21520f4736e3718e414f70366bea48b2df805b37a6a5193a1b', Something in that task is contaminated by target-specific value. Alex
Quoted 1 line · Alexander Kanavin
Markus, please. This is shotgun debugging
Of course it is, and these are things I wouldn't expect to fix the specific issue. But they would be useful anyway I thought
Quoted 3 lines · Alexander Kanavin
If you'd like to help, it
would be more productive to take the patch and build it locally to have first hand evidence of what is happening.
I try ;)
Had a look at core-image-sato and all icons were still present as far as i can tell And I tracked the issue down to this in allarch.bbclass
<https://git.yoctoproject.org/poky/tree/meta/classes-recipe/allarch.bbclass#n11>
Running addhandler allarch_package_arch_handler causes the diverge
Quoted 4 lines · Markus Volk
And I tracked the issue down to this in allarch.bbclass
https://git.yoctoproject.org/poky/tree/meta/classes-recipe/allarch.bbclass#n11Running addhandler
allarch_package_arch_handlercauses the diverge
How did you trace it? That doesn't make sense given what the test does, and the task where diverge begins.
Alex
Quoted 2 lines · Alexander Kanavin
How did you trace it? That doesn't make sense given what the test does, and the task where diverge begins.
First I ran
oe-selftest -r sstatetests.SStateHashSameSigs2.test_sstate_allarch_samesigs
2023-09-21 17:43:06,946 - oe-selftest - INFO - ---------------------------------------------------------------------- 2023-09-21 17:43:06,946 - oe-selftest - INFO - Ran 1 test in 68.893s 2023-09-21 17:43:06,946 - oe-selftest - INFO - FAILED 2023-09-21 17:43:06,946 - oe-selftest - INFO - (failures=1) 2023-09-21 17:43:11,370 - oe-selftest - INFO - RESULTS: 2023-09-21 17:43:11,370 - oe-selftest - INFO - RESULTS - sstatetests.SStateHashSameSigs2.test_sstate_allarch_samesigs: FAILED (66.80s) 2023-09-21 17:43:11,380 - oe-selftest - INFO - SUMMARY: 2023-09-21 17:43:11,380 - oe-selftest - INFO - oe-selftest () - Ran 1 test in 68.896s 2023-09-21 17:43:11,380 - oe-selftest - INFO - oe-selftest - FAIL - Required tests failed (successes=0, skipped=0, failures=1, errors=0)
to verify that I can reproduce the issue.
Then I looked at the code and couldn't find any suspicious activity in meson so I thought it may be an internal problem. I removed all non-essential stuff from the recipe
SUMMARY = "GTK+ icon theme"
DESCRIPTION = "The Adwaita icon theme is the default icon theme of the GNOME desktop \
This package package contains an icon theme for Gtk+ 3 applications." HOMEPAGE = "<https://gitlab.gnome.org/GNOME/adwaita-icon-theme>" BUGTRACKER = "<https://gitlab.gnome.org/GNOME/adwaita-icon-theme/issues>" SECTION = "x11/gnome"
LICENSE = "LGPL-3.0-only | CC-BY-SA-3.0"LIC_FILES_CHKSUM = "file://COPYING;md5=c84cac88e46fc07647ea07e6c24eeb7c <file://copying;md5=c84cac88e46fc07647ea07e6c24eeb7c/> \
file://COPYING_CCBYSA3;md5=96143d33de3a79321b1006c4e8ed07e7 <file://copying_ccbysa3;md5=96143d33de3a79321b1006c4e8ed07e7/> \
file://COPYING_LGPL;md5=e6a600fd5e1d9cbde2d983680233ad02 <file://copying_lgpl;md5=e6a600fd5e1d9cbde2d983680233ad02/>"
GNOMEBASEBUILDCLASS = "meson"GTKIC_VERSION = "4"
inherit gnomebase pkgconfig gtk-icon-cache allarch
SRC_URI[archive.sha256sum] = "2442bfb06f4e6cc95bf6e2682fdff98fa5eddc688751b9d6215c623cb4e42ff1" FILES:${PN}-doc = "${datadir}/licenses"
BBCLASSEXTEND = "native nativesdk"
but still the test failed. In the next step I excluded inherit allarch and the test succeeded
Looking at allarch.bbclass, this was the most suspicious part to me and first thing I tried was commenting out #addhandler allarch_package_arch_handler
2023-09-21 17:08:57,046 - oe-selftest - INFO - Adding: "include selftest.inc" in /home/flk/poky/build-core-st/conf/local.conf 2023-09-21 17:08:57,046 - oe-selftest - INFO - Adding: "include bblayers.inc" in bblayers.conf 2023-09-21 17:08:57,046 - oe-selftest - INFO - test_sstate_allarch_samesigs (sstatetests.SStateHashSameSigs2.test_sstate_allarch_samesigs) 2023-09-21 17:10:02,922 - oe-selftest - INFO - ... ok 2023-09-21 17:10:02,955 - oe-selftest - INFO - ---------------------------------------------------------------------- 2023-09-21 17:10:02,955 - oe-selftest - INFO - Ran 1 test in 68.008s 2023-09-21 17:10:02,955 - oe-selftest - INFO - OK 2023-09-21 17:10:07,700 - oe-selftest - INFO - RESULTS: 2023-09-21 17:10:07,700 - oe-selftest - INFO - RESULTS - sstatetests.SStateHashSameSigs2.test_sstate_allarch_samesigs: PASSED (65.88s) 2023-09-21 17:10:07,713 - oe-selftest - INFO - SUMMARY: 2023-09-21 17:10:07,713 - oe-selftest - INFO - oe-selftest () - Ran 1 test in 68.008s 2023-09-21 17:10:07,713 - oe-selftest - INFO - oe-selftest - OK - All required tests passed (successes=1, skipped=0, failures=0, errors=0)
Quoted 6 lines · Markus Volk
How did you trace it? That doesn't make sense given what the test does, and the task where diverge begins.
but still the test failed. In the next step I excluded inherit allarch and the test succeeded
Looking at allarch.bbclass, this was the most suspicious part to me and first thing I tried was commenting out #addhandler
allarch_package_arch_handler
I see, thanks. Unfortunately I think you went down the wrong path.
The recipe is meant to be allarch since it doesn't build anything for the target and only installs pre-built data which is identical on all targets. This improves sstate reuse for one thing. This worked fine until this version update, but after the version update the signature for the newly added task write_config (which I think comes from meson) differs depending on target. Something creeps in there which varies between targets, so you need to for example use bitbake-diffsigs to compare the signatures (the test prints what they are, and bitbake-diffsigs will tell you what changed).
Alex
The refmanual says about allarch:
Unlike some distro recipes (e.g. Debian), OpenEmbedded recipes that produce packages that depend on tunings through use of the RDEPENDS and TUNE_PKGARCH variables, should never be configured for all architectures using allarch. This is the case even if the recipes do not produce architecture-specific output.
gtk-icon-cache.bbclass adds RDEPENDS and that probably makes the package incompatible with allarch 'gtk_update_icon_cache' is run in meson automatically:
<https://gitlab.gnome.org/GNOME/adwaita-icon-theme/-/blob/master/meson.build?ref_type=heads#L89>
I patched meson to not run gtk_update_icon_cache and the test also passed.
But this brings me back to my first idea, that it might be the right thing to do to remove inherit allarch?
Quoted 24 lines · Alexander Kanavin
Quoted 9 lines · On Thu, 21 Sept 2023 at 17:51, Markus Volk <[email protected]…
How did you trace it? That doesn't make sense given what the test
does, and the task where diverge begins.but still the test failed. In the next step I excluded inherit
allarch and the test succeededLooking at allarch.bbclass, this was the most suspicious part to me
and first thing I tried was commenting out
#addhandlerallarch_package_arch_handlerI see, thanks. Unfortunately I think you went down the wrong path.
The recipe is meant to be allarch since it doesn't build anything for the target and only installs pre-built data which is identical on all targets. This improves sstate reuse for one thing. This worked fine until this version update, but after the version update the signature for the newly added task
write_config(which I think comes from meson) differs depending on target. Something creeps in there which varies between targets, so you need to for example use bitbake-diffsigs to compare the signatures (the test prints what they are, and bitbake-diffsigs will tell you what changed).Alex
Quoted 10 lines · Markus Volk
The refmanual says about allarch:
Unlike some distro recipes (e.g. Debian), OpenEmbedded recipes that produce packages that depend on tunings through use of the RDEPENDS andTUNE_PKGARCHvariables, should never be configured for all architectures using allarch. This is the case even if the recipes do not produce architecture-specific output.gtk-icon-cache.bbclass adds RDEPENDS and that probably makes the package incompatible with allarch '
gtk_update_icon_cache' is run in meson automatically:
https://gitlab.gnome.org/GNOME/adwaita-icon-theme/-/blob/master/meson.build?ref_type=heads#L89I patched meson to not run
gtk_update_icon_cacheand the test also passed.
But this brings me back to my first idea, that it might be the right thing to do to remove inherit allarch?
No. We are not dropping allarch. The recipe output is not architecture specific, and doesn't change depending on target, regardless of what host executables it needs at build time or whether or not it has target-specific things in RDEPENDS. There are plenty of other allarch recipes such as ca-certificates which use both build time native dependencies and run time dependencies which work fine and pass the test, and this is not special in any way.
Please read again what I said about task signatures because the problem with allarch compatibility is there.
Alex
Quoted 2 lines · Alexander Kanavin
Please read again what I said about task signatures because the problem with allarch compatibility is there.
I read it and also looked at the signatures. But it didn't tell me as much as it would have told you. At least it pointed me in the right direction. But I needed more research to get a handle on it. In my case, it was a combination of two problems. In the case of the original recipe, it's just the first one.
- allarch must be last in inherit
'inherit gnomebase gtk-icon-cache allarch'
- you should not set GTK_IC = "4", as I had previously assumed
Quoted 7 lines · Markus Volk
Quoted 1 line · Alexander Kanavin
Please read again what I said about task signatures because the problem with allarch compatibility is there.
I read it and also looked at the signatures. But it didn't tell me as much as it would have told you. At least it pointed me in the right direction. But I needed more research to get a handle on it. In my case, it was a combination of two problems. In the case of the original recipe, it's just the first one.
- allarch must be last in inherit
'inherit gnomebase gtk-icon-cache allarch'
There are same name functions qemu_wrapper_cmdline() in allarch.bbclass and qemu.bbclass which is involved by meson.bbclass.
In qemu_wrapper_cmdline() of qemu.bbclass, it uses var ${OLDEST_KERNEL} which becomes a dependency:
63 QEMU_OPTIONS = "-r ${OLDEST_KERNEL} ${@d.getVar("QEMU_EXTRAOPTIONS_%s" % d.getVar('PACKAGE_ARCH')) or ""}"
But in qemu_wrapper_cmdline() of allarch.bbclass, the function has been set empty. So there is no such dependency and the task checksum don't change.
That's why inherit allarch after gnomebase which inherit meson doesn't fail the oe-test.
Many thanks.
Kai
Quoted 4 lines
- you should not set
GTK_IC= "4", as I had previously assumed
Signature · 2 lines
Kai Kang Wind River Linux