@jstancek is this OK as is, or would you like to include the alternate .net/jira/browse/xyz as well?
Derek Barbosa (b297826b) at 17 Mar 05:01
Merge branch 'ft-unowned-files' into 'main'
... and 1 more commit
Update the include list to reflect the current state of the maintained codebase and reduce the occurrence of "Unowned Files" label in gitlab. Also adjust the jiraComponent and SST fields for the LIB subcomponent.
Signed-off-by: Rafael Aquini [email protected]
Restartable sequence is a kernel feature that enables userspace code (like glibc) to handle percpu variables and even implement userspace spinlocks. It complements futex that enables userspace mutex. Add that as part of the locking component.
Signed-off-by: Waiman Long [email protected]
Derek Barbosa (c328cd96) at 16 Mar 13:25
Merge branch 'longman_rseq' into 'main'
... and 1 more commit
Derek Barbosa (74d5dceb) at 14 Mar 06:12
webhook: dochook: Improve logging
Derek Barbosa (92668b10) at 13 Mar 19:09
webhook: dochook: Improve logging
Derek Barbosa (2308ffd1) at 13 Mar 18:56
owners-schema: initalize the fallback field
I think I will remove the ALIAS portion of this MR and push v2 without it. Having this being gated on successful changes to cki_lib is a bit out of scope for trying to amend the logging situation.
TLDR cki deployments utilizing the entrypoint script still default to plain.
@tales-aparecida that's what I get for not git pull-ing!
thanks for the heads up.
I do welcome any and all comments
Yeah you have a point. All I know is that is what came up when I was grepping against cki_lib. In the next version of this MR, I can refactor and just scrap it entirely.
Frankly, I forget why the original code had this mapping stored in its own codebase rather than owners.
@ptalbert I reached out to Jarod last night before logging off, and it seems like the example you provided exists in cki_lib -- doing a yaml.load('FOO_VAR','{}')
I don't hate your approach. Would a good middle ground be just keeping stuff in utils/ or webhook/utils?
I was under the assumption the default is just plain?
cki_entrypoint.sh
function log() {
local timestamp
case "${CKI_LOGGING_FORMAT:-plain}" in
json|pack)
timestamp=$(date +%s.%N)
echo '{"timestamp": "'"${timestamp}"'", "logger_name": "cki_lib.entrypoint", "logger_level": "INFO", "_entry": "'"$1"'"}'
;;
*)
timestamp=$(date -u +%FT%T.%6N)
echo "${timestamp} - [INFO] - cki_lib.entrypoint - $1"
;;
esac
}
logger.py
logging_format = os.environ.get('CKI_LOGGING_FORMAT', 'plain')
I am probably reading that bashism totally wrong though.
But there is always the option for users to revert to plain, as in the webhook docs. Let me give it a few more runs locally, last time I looked it wasnt that bad
by this you mean the EMAIL aliases or setting the environment vars?
Derek Barbosa (eff79a85) at 12 Mar 18:50
Merge: chores: readme: add packages used by the linting container
... and 1 more commit
JIRA: https://issues.redhat.com/browse/RHELMISC-28328
Also, add a more specific path to ruff for ignoring site packages
Signed-off-by: Derek Barbosa [email protected]