linux/poky.git/meta/classes, branch kirkstone-4.0.22

populate_sdk_base: inherit nopackages

2024-10-07T12:43:22+00:00

Since this bbclass sets PACKAGES = "", inherit the nopackages class to skip the various packaging functions which wouldn't do anything anyway. This fixes errors from buildhistory changes where packages-split would be empty. e.g. meta-toolchain build now fails with: | DEBUG: Executing shell function buildhistory_list_pkg_files | find: ".../meta-toolchain/1.0/packages-split/*": No such file or directory | WARNING: exit code 1 from a shell command. | DEBUG: Python function buildhistory_emit_pkghistory finished (From OE-Core rev: 3b646f322b4ffd5ed520f3815ce0726cf225ced2) Signed-off-by: Martin Jansa Signed-off-by: Richard Purdie Signed-off-by: Atharva Nandanwar Signed-off-by: Steve Sakoman

testexport: fallback for empty IMAGE_LINK_NAME

2024-09-27T12:57:35+00:00

if IMAGE_LINK_NAME is set empty to disable the symlinking for image artifacts in deploy, testexport fails, as the path assembly is incorrect. In that case fallback to IMAGE_NAME (From OE-Core rev: 0c1d098e6dd08fa3a5aafca656457ac6badcef89) (From OE-Core rev: a821a5810ae6edba5882a43e3c60c142f2acda9d) Signed-off-by: Konrad Weihmann Signed-off-by: Richard Purdie Signed-off-by: Steve Sakoman

testimage: fallback for empty IMAGE_LINK_NAME

2024-09-27T12:57:35+00:00

if IMAGE_LINK_NAME is set empty to disable the symlinking for image artifacts in deploy, testimage fails, as the path assembly is incorrect. In that case fallback to IMAGE_NAME (From OE-Core rev: c7a4e7e294992acc589c62adcaf6cd32659f2f9b) (From OE-Core rev: 5b4c7a39f7a99d5c98eab3aaf693f4d3000c7ac1) Signed-off-by: Konrad Weihmann Signed-off-by: Richard Purdie Signed-off-by: Steve Sakoman

buildhistory: Simplify intercept call sites and drop SSTATEPOSTINSTFUNC usage

2024-09-27T12:57:35+00:00

We planned to drop SSTATEPOSTINSTFUNC some time ago with the introduction of postfuncs. Finally get around to doing that which should make the buildhistory code a little more readable. Unfortunately ordering the buildhistory function calls after the sstate ones is difficult without coding that into the sstate class. This patch does that to ensure everything functions as expected until we can find a better way. This is still likely preferable than the generic sstate postfuncs support since the function flow is much more readable. (From OE-Core rev: 78ca086441b21dedd9c471a3d3200c24fd9ec8d2) Signed-off-by: Richard Purdie (cherry picked from commit c9e2a8fa2f0305ef1247ec405555612326f798f8) Signed-off-by: Steve Sakoman

buildhistory: Restoring files from preserve list

2024-09-27T12:57:35+00:00

This fix will ensure that, when we activate feature `BUILDHISTORY_RESET`, files marked to keep on feature `BUILDHISTORY_PRESERVE` will indeed exist is buildhistory final path since they are moved to buildhistory/old but not restored at any point. (From OE-Core rev: 8160fd3c042283a47a2601e1797847c303e1e7e5) Signed-off-by: Pedro Ferreira Signed-off-by: Richard Purdie (cherry picked from commit 9f68a45aa238ae5fcdfaca71ba0e7015e9cb720e) Signed-off-by: Steve Sakoman

buildhistory: Fix intermittent package file list creation

2024-09-27T12:57:35+00:00

The directory that buildhistory_list_pkg_files writes to during do_package is created by do_packagedata so a clean buildhistory doesn't have files-in-package written during the first build since packagedata happens after do_package. Ensure the output package folder is created to avoid missing files-in-package.txt files. Also it ensures that in case of `find` fails we leave with a hard error instead of hiding the error on the for loop. (From OE-Core rev: 6817b012763fc32cdcffe30163a304da3ed59ae1) Signed-off-by: Pedro Silva Ferreira Signed-off-by: Richard Purdie (cherry picked from commit 8de9b8c1e199896b9a7bc5ed64967c6bfbf84bea) Signed-off-by: Steve Sakoman

image_types.bbclass: Use --force also with lz4,lzop

2024-08-28T13:49:22+00:00

Several conversion commands already make use of 'force' option in the compression, which enables overwriting existing files without prompting. Since occasionally an existing residual destination file from a previously aborted or failed task can prevent the re-execution of the conversion command task, by enabling the 'force' option also for lz4 and lzop compression commands we can avoid following kind of BitBake failures with these compressors: | DEBUG: Executing shell function do_image_cpio | 117685 blocks | 2 blocks | example-image.cpio.lz4 already exists; do you want to overwrite (y/N) ? not overwritten | Error 20 : example-image.cpio : open file error | WARNING: exit code 20 from a shell command. ERROR: Task (.../recipes-core/images/example-image.bb:do_image_cpio) failed with exit code '1' (From OE-Core rev: 623ab22434909f10aaf613cd3032cc2a2c6e3ff9) (From OE-Core rev: 35a92023766b562dab9f8561cc3adb4938b03692) Signed-off-by: Niko Mauno Signed-off-by: Richard Purdie Signed-off-by: Steve Sakoman

uboot-sign: Fix index error in concat_dtb_helper() with multiple configs

2024-07-24T14:51:58+00:00

Commit 60774248a5570899a66f9f88e597cc4f723d6278 solves issues when i or j have been set in other shell functions. Since j is not unset after the inner loop, the image will only be copied for the first configuration. Unsetting i and j after the loops also prevents index issues in other functions. (From OE-Core rev: f6de96c9fa8d0b6c81c32016f342ad93c8940d9e) Signed-off-by: Florian Amstutz Signed-off-by: Steve Sakoman

linuxloader: add -armhf on arm only for TARGET_FPU 'hard'

2024-07-09T13:06:09+00:00

There are two types of soft FPU options for arm, soft and softfp, and if using the latter the wrong dynamic loader will be used. E.g. go will link against ld-linux-armhf.so.3, but libc6 will only ship a ld-linux.so.3, so go programs will fail to start. Fix this by instead checking for TARGET_FPU being 'hard' and then applying the suffix. (From OE-Core rev: c7426629245db2ea8d9f3cf25b575ac31b5a83b0) Signed-off-by: Jonas Gorski Signed-off-by: Richard Purdie (cherry picked from commit 07b4c7a2bd23f8645810e13439e814caaaf9cd94) Signed-off-by: Steve Sakoman

classes: go-mod: do not pack go mod cache

2024-05-29T12:24:08+00:00

Clean go module cache from builddir to prevent it of beeing packed. (From OE-Core rev: c850931590ff22da4d38756f957b88e04078c76c) Signed-off-by: Stefan Herbrechtsmeier Signed-off-by: Lukas Funke Signed-off-by: Alexandre Belloni (cherry picked from commit 328bea56dec8f83b5c118f567e122510f9243087) Signed-off-by: Jose Quaresma Signed-off-by: Steve Sakoman