bitbake: Merge latest upstream#47
Draft
Shreejit-03 wants to merge 161 commits intoni:nilrt/master/nextfrom
Draft
bitbake: Merge latest upstream#47Shreejit-03 wants to merge 161 commits intoni:nilrt/master/nextfrom
Shreejit-03 wants to merge 161 commits intoni:nilrt/master/nextfrom
Conversation
Signed-off-by: Jörg Sommer <joerg.sommer@navimatix.de> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The regex used to detect bitbake files was improved in patch 9.1.1732. Update the plugin to match that logic and fix a bug where the wrong expression was used to set the file type Signed-off-by: Joshua Watt <JPEWhacker@gmail.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
…S_BUILTIN The actual variable name used in oe-core is OE_FRAGMENTS_BUILTIN, which is in a consistent format with other OE_FRAGMENTS_XXX vars. Signed-off-by: Chen Qi <Qi.Chen@windriver.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
…ment --global was defined twice (for settings command as a whole, and specifically for set/unset sub-commands), which caused 'settings --global set ...' to be allowed, and at the same time not have any effect. 'settings set --global' is the only valid form after this change. Signed-off-by: Alexander Kanavin <alex@linutronix.de> Tested-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Describe the roles of AUTOREV and SRCREV_FORMAT in the fetching step. Add their missing descriptions to the variable reference section. Fixes [YOCTO #14498] Signed-off-by: Adam Blank <adam.blank.g@gmail.com> Reviewed-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Fix encountered style issues in the fetching section. Mention that "subdir" and fetcher specific parameters can stack-up. Signed-off-by: Adam Blank <adam.blank.g@gmail.com> Reviewed-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Current wording makes it difficult to understand what exactly is done with INHERIT-ed anonymous functions and when. Signed-off-by: Adam Blank <adam.blank.g@gmail.com> Reviewed-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
… the proto in the message Signed-off-by: Chris Laplante <chris.laplante@agilent.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Adding documentation for having an empty first component for the SRCREV_FORMAT. Signed-off-by: Patrick Vogelaar <patrick.vogelaar.dev@mailbox.org> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Extend the current documentation of PERSISTENT_DIR: - Mention that it also contains the hash equivalence database when BB_HASHSERVE = auto. - Mention that it should not be shared between builds. [YOCTO #15727] Signed-off-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
After a terminology review by Antonin we brainstormed something less confusing than 'build' and 'build directory' for the place where bitbake-setup clones layers and creates a *bitbake* build directory in. People are bound to get these two confused and mix them up, and 'setup' is much more distinct and aligns nicely with 'bitbake-setup'. It's also not claimed by anything else in OE/Yocto. So before: top-dir -> build-dir (can be several) -> bitbake build dir, layer dir, config dir Now: top-dir -> setup-dir (can be several) -> bitbake build dir, layer dir, config dir This also updates the respective command line options, I understand it's a breaking change, but as before the tweaks are simple and we need to get the terminology right for the users, and now is the time to do it. Signed-off-by: Alexander Kanavin <alex@linutronix.de> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This adds support for the new bitbake-setup:
* Maintain backwards compatibility for poky-classic releases
* Assume bitbake-setup for all releases 'master' and forward
* When no ".templateconf" assume Poky and not OE
* Generate and use "init-build-env" script instead of 'oe-init-build-env'
directly, but "cd" to the clone and not the local installation.
* Include "bblayers.conf" because "cookerdata.py" now explicitly
required it, and let it be over-ridden by the usual "toaster-bblayers.conf".
* Force the value "TOPDIR" in "toaster.conf" to the local project's directory.
Normally "cookerdata.py" sets TOPDIR from getcwd(), but somehow in bitbake-setup
mode the CWD at this point has become ".../layers/openembedded-core", so
"bitbake.conf and all other files cannot be found. This non-obvious problem was
hard to discover. Until this is understood, the forced value will make things work.
[YOCTO #16012]
Signed-off-by: David Reyna <David.Reyna@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
…tings_path' Rename the function to align with the corresponding 'global_settings_path' function, and move it down just above the later. This is done to differentiate from the built-in default settings, and to align with the other places in the code that use 'topdir_settings' (or 'global_settings') Signed-off-by: Johannes Schneider <johannes.schneider@leica-geosystems.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Setting a key=value in a new section would raise a key-error when
using in-memory settings, e.g:
./bin/bitbake-setup --setting default top-dir-prefix /tmp/bitbake-setup --setting default top-dir-name gs settings foo bar baz
Loading settings from
/tmp/bitbake-setup/gs/settings.conf
Traceback (most recent call last):
File "/tmp/bitbake-setup/bitbake/./bin/bitbake-setup", line 853, in <module>
main()
File "/tmp/bitbake-setup/bitbake/./bin/bitbake-setup", line 838, in main
all_settings = merge_settings(builtin_settings, global_settings, topdir_settings, args.setting)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/tmp/bitbake-setup/bitbake/./bin/bitbake-setup", line 733, in merge_settings
all_settings[section][setting] = value
~~~~~~~~~~~~^^^^^^^^^
KeyError: 'foo'
Signed-off-by: Johannes Schneider <johannes.schneider@leica-geosystems.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Rename the function to reflect which (of the possibly many) files in the build/config directory it writes to. Signed-off-by: Johannes Schneider <johannes.schneider@leica-geosystems.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Construct the 'src_uri' separately, and then pass either variant into one call that creates the Fetch. Making use of format-strings to shorten/simplify the code. Also: using 'proto' instead of 'type' for a variable name, to avoid the protected keyword. Signed-off-by: Johannes Schneider <johannes.schneider@leica-geosystems.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
All places that called into 'check_setupdir_files' did the same preparation step to load the upstream-config.json and then pass it into the function. Since the 'setuppath' is already passed into the function, and the name and relative location of the upstream-config.json is fixed, constructing the file path and loading the json could be done in the function. De-duplicate code by loading the json inside the function instead. Signed-off-by: Johannes Schneider <johannes.schneider@leica-geosystems.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When initializing a build setup from a conf.json that only sets 'rev' to a tag or branch, the actual revision would not be captured or logged. To capture the current layer state after an 'init' or 'update', the checkout_layers function is extended to store the revision the bb.fetch.Fetch pulled, and write that information into a sources-fixed-revisions.json file. This file can then be fed back into bitbake-setup init as: --sources-overrides This new 'sources-fixed-revisions.json' is written during 'update_build' and stored alongside the 'config-upstream.json' in the config dir. And put with the later under version control by calling 'commit_config" after 'update_build'. The use of 'deepcopy' is necessary to not modify the original input data - which python passes around as reference. Signed-off-by: Johannes Schneider <johannes.schneider@leica-geosystems.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
…tive
{THISDIR} is a special value token that can be used in the list of enabled
layers to specify the layer location relative to the confguration file:
https://git.openembedded.org/bitbake/commit/?id=b3153be29de8b8570b0c184369bd41f4c646cf92
This replaces the token with an explicit separate keyword for such layers:
so that special processing to determine the final value can be avoided, and
the feature can be formalized in the json schema:
instead of
"bb-layers": [
"{THISDIR}/meta-my-project"
]
this allows
"bb-layers-relative": [
"meta-my-project"
Going forward I think we should strive to avoid any further special value tokens.
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The difference between bb-layers and bb-layers-relative is unclear as both are relative paths. Rename one to "file-relative" which makes it clear it is relative to the current file, without becomming a long name. https://lists.openembedded.org/g/bitbake-devel/message/18296 Based on a patch from Alexander Kanavin <alex@linutronix.de> but with different naming. Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
For use in the documentation. Signed-off-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add documentation for bitbake-setup. This is split in 6 sections, including a quick start guide, terminology and reference. Signed-off-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
As of now, the variable history for builtin fragments looks like this (edited for clarity): $ bitbake-getvar MACHINE # # $MACHINE [2 operations] # set ast.py:368 [check_and_set_builtin_fragment] # "qemux86-64" # pre-expansion value: # "qemux86-64" MACHINE="qemux86-64" User can't know where MACHINE was set, this is bad. This patch tries to reconstruct a MACHINE history from OE_FRAGMENTS history. With this patch, history looks like this (for a simple case): $ bitbake-getvar MACHINE NOTE: Starting bitbake server... # # $MACHINE [2 operations] # set .../auto.conf:2 # "qemux86-64 (OE_FRAGMENTS contains "machine/qemux86-64")" # pre-expansion value: # "qemux86-64" MACHINE="qemux86-64" The path where the "machine/qemux86-64" fragment was added to OE_FRAGMENTS is displayed, this is definitely better. Fixes [YOCTO #15939] Signed-off-by: Yoann Congal <yoann.congal@smile.fr> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
While it can be argued how useful a progressbar created with 0 as maximum value is, it should still show two states, started (empty) and finished (full). Setting the maxval to _DEFAULT_MAXVAL instead will accomplish this. Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Ensure that every time a lockfile is created, it is also unlocked in a finally: block. Otherwise, the lock file remains open and Python warns: ResourceWarning: unclosed file <_io.FileIO name='/tmp/bitbake-fetch-srfv29ek/download/git2/git.openembedded.org.bitbake.lock' mode='ab+' closefd=True> Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Python 3.14 complains about these:
Traceback (most recent call last):
File "/mnt2/zozo/yocto-5.3/bitbake/lib/bb/ui/knotty.py", line 722, in main
termfilter.updateFooter()
~~~~~~~~~~~~~~~~~~~~~~~^^
File "/mnt2/zozo/yocto-5.3/bitbake/lib/bb/ui/knotty.py", line 339, in updateFooter
lines = self.getlines(content)
File "/mnt2/zozo/yocto-5.3/bitbake/lib/bb/ui/knotty.py", line 370, in getlines
lines = lines + 1 + int(len(line) / (self.columns + 1))
~~~~~~~~~~~~~^~~
TypeError: can only concatenate str (not "int") to str
and
Traceback (most recent call last):
File "/mnt2/zozo/yocto-5.3/bitbake/lib/bb/ui/knotty.py", line 722, in main
termfilter.updateFooter()
~~~~~~~~~~~~~~~~~~~~~~~^^
File "/mnt2/zozo/yocto-5.3/bitbake/lib/bb/ui/knotty.py", line 341, in updateFooter
for tasknum, task in enumerate(tasks[:(self.rows - 1 - lines)]):
~~~~~~~~~~^~~
TypeError: unsupported operand type(s) for -: 'str' and 'int'
Make sure getting the number of rows and columns from the terminal
via the environment variables LINES and COLUMNS are returned as a
pair of integers. This matches the return value of ioctl_GWINSZ().
Signed-off-by: Zoltán Böszörményi <zboszor@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The symlink variable is never defined, and it appears that it should have been siteconf_symlink. Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When the user is asked to select a configuration, the list is zero-indexed: Distribution configuration variants: 0. distro/poky 1. distro/poky-altcfg 2. distro/poky-tiny It is more user friendly to start the lists from "1". Signed-off-by: Ross Burton <ross.burton@arm.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
…tings top-dir-prefix and top-dir-name only make sense to be set in the global settings file, so if they are specified force them to the global settings file, and warn the user that this is happening. Signed-off-by: Joshua Watt <JPEWhacker@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The poky distro is now stripped of its repositories. Toaster had used that
repo as both a quick way to get the three core layers (openembedded-code,
meta-poky, and meta-yocto-bsp) but also a way to get the related bitbake,
which was important because bitbake is not in the Layer Index. Now, Toaster
must explicitly clone each of the core layers plus the proper bitbake layer.
Notes on "orm/fixtures":
* The "orm/fixtures" files used to hard code the three core layers to
point to the poky distro, ignoring the Layer Index values
* The script "lsupdates.py" is now allowed to use the Layer Index to
properly initialize these layers.
* The fixture files are still used because (a) they define the supported
releases for Toaster, plus (b) they now define the proper bitbake repo
path and branches to associate with the YP releases.
Notes on "localhostbecontroller.py":
* Always insert the proper bitbake into the layer clone list
* Insure that both 'BBBASEDIR' and 'BITBAKEDIR' are defined in the
sub-shells so that oe-init finds the right bitbake (e.g.
'oe-buildenv-internal')
[YOCTO #16012]
Signed-off-by: David Reyna <David.Reyna@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The property 'uri' is a simplification of the property 'remotes'. It is used to provide only one URI. Signed-off-by: Corentin Guillevic <corentin.guillevic@smile.fr> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Copied from bitbake-server/worker/layers. This is seen also for bitbake command invocations. Signed-off-by: Peter Marko <peter.marko@siemens.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
… variable This fixes running 'bitbake-setup init' in an active bitbake session when a top dir settings file is absent. In this case, bitbake-setup was trying to make a whole new top directory under the current directory. It will will now use the existing top directory, deducing it from BBPATH (similar to status and update operations), which was always the intent. The code was relying on an existence of a top directory settings file, which was always present in earlier prototypes, but is currently created only when a setting is explicitly written into it. This logic is changed to use the bitbake-setup's cache directory, which is also explicitly created. Signed-off-by: Alexander Kanavin <alex@linutronix.de> Signed-off-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
…e repo otherwise If bitbake-setup is packaged and obtained from pypi, or another package feed it needs to be able to find standard configurations. It is not impossible to package configurations into some location inside the package and add logic to find them, but it is much easier to just fetch the bitbake repo and use that as a configuration registry. The other benefit is that such packaged/installed bitbake-setup will not have to be updated in order to obtain latest configurations: it will simply check the registry and fetch them as needed, which is the original intent of config registries as repos. Signed-off-by: Alexander Kanavin <alex@linutronix.de> Signed-off-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The goal for the 'list' command is to not leave any files or directories behind, as it may be issued before the user has decided where to place the top directory. 'list' redirects the top directory to a temporary location, but it didn't do the same for the downloads, which led to the default top directory being polluted with them. This has now been fixed. Signed-off-by: Alexander Kanavin <alex@linutronix.de> Signed-off-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In those cases when the user enters a URL that points to the html representation of a json file, and not the actual json file, we can at least let them know that the URL point to a file which is not a json. It will help when a user copy and paste a URL from a browser and the tool cannot parse it as json. Signed-off-by: Daiane Angolini <daiane.angolini@foundries.io> Signed-off-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
…not the whole settings This will help the following commit. Signed-off-by: Alexander Kanavin <alex@linutronix.de> Signed-off-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
status/update operations should be using the registry that was used in initializing a setup, but they've been using the registry from settings. This is now corrected. Signed-off-by: Alexander Kanavin <alex@linutronix.de> Signed-off-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
… configuration This was done before setting up a new bitbake configuration, so that it does not write over the old one, comparing the old and the new bitbake configurations, and taking one of three possible actions: - no changes, no need to keep a backup; - changes exist, do not replace existing config - changes exist, replace existing config. All three options correctly print the eventual outcome, so this early print is unnecessary and confusing. Signed-off-by: Alexander Kanavin <alex@linutronix.de> Signed-off-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The order was non-deterministic which did not help finding the number for the right one in the list. Signed-off-by: Alexander Kanavin <alex@linutronix.de> Signed-off-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Alexander Kanavin <alex@linutronix.de> Signed-off-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This prepares the data for the next commit. Signed-off-by: Alexander Kanavin <alex@linutronix.de> Signed-off-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Change some terms to use more specific wording: Generic Configuration -> Configuration Template Specific Configuration -> Configuration Instance Signed-off-by: Daiane Angolini <daiane.angolini@foundries.io> Reviewed-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Make the output to match the terminology used in the documentation for some terms. Signed-off-by: Daiane Angolini <daiane.angolini@foundries.io> Reviewed-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If we use the default of bitbake-setut init we may not create anything (as the first default for an empty directory is N). If user does not chose a number, the script asks again for a number. Signed-off-by: Daiane Angolini <daiane.angolini@foundries.io> Reviewed-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If sources is empty, nothing happens. So let's just not say it's optional, and let's say the minimum is bitbake source code. Signed-off-by: Daiane Angolini <daiane.angolini@foundries.io> Reviewed-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Make it possible to give fragments descriptions as it can help users determine which fragments to pick during bitbake-setup init. It is still possible to provide a list (with no descriptions) for backwards-compatibility. Update the documentation accordingly. Signed-off-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If unset, the existing behavior is preseved. The use case is sharing the database in bitbake-setup's driven builds without having to set up/start/stop a common single server shared between them (this is added to bitbake-setup in the next commit). Also create the specified directory if it doesn't yet exist. Add a check that the directory is not on a NFS path, and error out then, guiding the user to set up a standalone hash equivalency server. Note: the check runs 'stat' executable with a format string parameter, ensuring it prints only the filesystem type and nothing else. Python's standard library does not have a statfs() wrapper, and using ctypes to call into it would've required delicate, crash-prone data type definitions, full of magic numbers. Signed-off-by: Alexander Kanavin <alex@linutronix.de> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
…/add layer
While develop a registry configuration adding a new layer and then back
to previous configuration without the _has_local_modifications breaks
because the code expects the layer to be in sources-fixed-revisions.json.
This can be reproduced with next steps:
- Call bitbake-setup with configX.
- Remove a LayerY from the configX.
- Call bitbake-setup with configX.
- Add the layerY again to configX.
- Call bitbake-setup with configX (break).
Fixes:
```
File "/workspaces/ls/meta-ls/bitbake/bin/bitbake-setup", line 1137, in <module>
main()
~~~~^^
File "/workspaces/ls/meta-ls/bitbake/bin/bitbake-setup", line 1130, in main
args.func(top_dir, all_settings, args, d)
~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspaces/ls/meta-ls/bitbake/bin/bitbake-setup", line 735, in build_update
build_status(top_dir, settings, args, d, update=True)
~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspaces/ls/meta-ls/bitbake/bin/bitbake-setup", line 719, in build_status
update_build(new_upstream_config, confdir, setupdir, layerdir, d,
~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
update_bb_conf=args.update_bb_conf)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspaces/ls/meta-ls/bitbake/bin/bitbake-setup", line 408, in update_build
sources_fixed_revisions = checkout_layers(layer_config, confdir, layerdir, d)
File "/workspaces/ls/meta-ls/bitbake/bin/bitbake-setup", line 184, in checkout_layers
elif _has_local_modifications(r_name, repodir_path):
~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^
File "/workspaces/ls/meta-ls/bitbake/bin/bitbake-setup", line 143, in _has_local_modifications
rev = fixed_revisions['sources'][r_name]['git-remote']['rev']
~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^
KeyError: 'meta-freescale'
```
Signed-off-by: Anibal Limon <anibal@limonsoftware.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The recipe name matching wasn't fully explained and were missing a few important corner cases such as recipes without a version, or using the % character outside of the version part of the filename for example. This adds the last missing bits to cover every append naming scenario that exist today. This was derived from lib/bb/cooker.py:get_file_appends(). Reported-by: Robert P. J. Day <rpjday@crashcourse.ca> Signed-off-by: Quentin Schulz <quentin.schulz@cherry.de> Reviewed-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Replace fragile parsing of 'git branch' output with the more reliable 'git name-rev --name-only HEAD' command to get the current branch name. This avoids parsing issues and is the recommended way to get the current branch name. Signed-off-by: Osama Abdelkader <osama.abdelkader@gmail.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Configs from registry, and configs on local filesystem were tested, but not configs obtained over http URIs. Signed-off-by: Alexander Kanavin <alex@linutronix.de> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There were no tests of oe-fragments-one-of so modify gizmo-notemplate to gizmo-notemplate-fragments-one-of, and test both formats for the options (with or without a description). In get_setup_path(), if the configuration specified on the command-line contains fragments, join everything with "-" as this is what will be the final directory name of the setup. This does not affect other current variants. Signed-off-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add the message filtering directly in the python module where the error happens, so we do not need to repeat the filter in all consumers. This will also make sure the filter is applied in any user of createDaemon(). Fixes [YOCTO #16080] Suggested-by: Chen Qi <Qi.Chen@windriver.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Give descriptions to the fragment choices to help the user know what to pick during 'bitbake-setup init'. These descriptions were based on the descriptions found in the corresponding configuration files in OE-Core/meta-yocto. Also while we're at it, s/Distributions/distributions/ as "machines" is lowercase in "Target machines" above. Signed-off-by: Antonin Godard <antonin.godard@bootlin.com> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If user add "" to dl-dir as DL_DIR, it will show FetchError as following.
raise FetchError("Unable to fetch URL from any source.", u)
bb.fetch2.FetchError: Fetcher failure for URL: 'git://git.openembedded.org/bitbake;protocol=https;rev=master;branch=master;destsuffix=bitbake'. Unable to fetch URL from any source.
But from this log, user can't locate the real reason which is that the path "/aa/bb" is invalid.
So I make this fix, even if user add "" to dl-dir, the dl-dir will be the valid path.
Other settings value also need this fix.
Signed-off-by: Liu Yiding <liuyd.fnst@fujitsu.com>
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We see errors on the autobuilder which imply we migtht be running out of file descrptors. Add some debugging to try and determine if this really is the case. For example the garbage collection may not be running quickly enough to keep up with the code. Count up the open files from /proc/self/fd, list them and then also try and run lsof (which might fail if no file descriptors are available). Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Bazaar has been of limited use outside of Launchpad for a while. Most uses within launchpad now use git. The original bazaar client was also replaced by breezy, which whilst supporting the bazaar repositories, mostly works with git as the backend now too. The existing code has been broken for a while and nobody seems to be using it, lets remove the obsolete technology. Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The OpenSUSE Build Service fetcher was for a specialist niche usecase which never went into production use as far as I know. At this point it makes sense to remove it as it is unused and bitrotted. Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
CVS is extremely dated technology and is not used anywhere significant/public any more. The existing code is not well maintained and is hard to test without any public CVS repositories. To be honest, if you're still using CVS you really shouldn't be. Remove the fetcher as obsolete and a generally bad idea now. Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
20 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Merge latest from upstream. No conflicts.
AB#3599972