given a task ID, fetch its details and display them to a table or to
json with `-o json`. Table view also adds a calculated task duration and
can optionally include the submitter information as well as a link to
logs for the task.
* --no-color - to skip colorizing output
* --L|--logs - include column with link to logs
* --submitter - show submitter
* --no-wait - control whether to wait until a task completes to output
(table mode)
`peridot task logs` will accept either a build ID or package name. In
the latter case, the project will be queried for the latest build for
the given package name. With the build task ID in hand, the task can be
queried for its BUILD_ARCH subtasks, which contain the logs for the
binary RPM builds.
Passing `-C` or `--cwd` with a directory will change into that directory
before writing files. That directory **must** exist beforehand.
Otherwise, $PWD is used.
By default, a separate file will be created for each
<task>-<subtask>-<architecture> tuple, which accounts for modular packages
with multiple outputs.
Passing the `-c` or `--combined` flag will combine all the logs into a
single file, named <task id>.log
Currently the signing process assumes that the "Package" size stays
consistent post-build and post-sign. This is a small
backwards-compatible change that properly sets a post-sign size in the
metadata during sync.
Go 1.22.2 fixes some CVEs so we should upgrade. This PR reverts support
for "ppc64" in rules_go. It was a recent addition but breaks builds for
ppc64le. @platforms recently added explicit ppc64le support so when
rules_go switches to that we can remove this patch
UI was upgraded for hydra-client v2.0.2, but the chart deployment is
still deploying an incompatible server version causing a deployment
mismatch and failure to deploy the new peridot frontend
Currently the Rust builds fails, most likely due to a lock that happens
with aarch64. Either the heartbeat tick dies early and timeout happens
right before success or something else makes heartbeats not work at all.
Best solution for now is to just increase timeout.
This endpoint doesn't ask for a file, so this is superfluous and,
therefore, confusing.. me. It's confusing me.
Signed-off-by: Neil Hanlon <neil@rockylinux.org>