lp:~apw/+git/kteam-tools

Author: Andy Whitcroft
Author Date: 2023-04-17 21:29:30 UTC

swm: prepare-packages -- block based on unprepared underspins

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-09-22 10:01:33 UTC

mainline-builds: add schroot breakage

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-09-16 21:31:20 UTC

swm: bug/flags -- switch all users of flags to the flags collection

Switch over legacy users of flags to the share flags collection.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-07-13 08:23:46 UTC

swm-jira-sync: checkout and update kernel-users

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-06-29 11:46:21 UTC

swm: package -- ready_for_testing should pass when in later pockets

In a partial respin the kernel is still ready_for_testing of some of it
is in -proposed and other packages are in -updates or -security. Handle
split occupancy when considering our ready_for_testing status.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-06-16 20:07:49 UTC

swm: automated-testing -- follow adt-matrix move to PS5

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-09-23 09:12:00 UTC

swm: package/pocket_clear -- reset found for each package

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-09-20 10:23:47 UTC

swm: manager -- clarify crank logging for closed trackers

When a tracker is closed it may or may not be purgable. Clarify the
diagnostics for this case.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-11-10 08:23:19 UTC

swm: stop sending messages to decommissioned rabbitmq servers

We have been duplicating testing messages into the old and new rabbitmq
servers. The old has now been decomissioned, stop sending messages to
it. Also stop attempting to send and quietly failing to send messages
to "shankbot" which has long been subsumed by swm itself.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-09-14 21:00:52 UTC

ancillary-bot: ancillary-lookup -- report bad PPA references

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-09-14 19:01:06 UTC

ktl/kernel_series: package.feeder is deprecated and no longer used

We have deprecated package.feeder in favour of package.depends but
retained an alias for live users. All of those are now updated so we
can now drop this support.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-09-09 21:36:10 UTC

swm: Secrets -- handle config=None correctly

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-08-04 10:08:26 UTC

swm-jira-sync: fix logging to include (noop) for issue tag updates

Seems we are calculating whether our tag operations are needed
and whether they have already been performed but only logging the
former. Fix the logging format to include the latter.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-08-04 09:12:57 UTC

ktl/msgq: drop deprecated rabbitmq servers

We have now consolidated all services onto the new PS5 rabbitmq
instance. Both argyle and prod-kernel-testing PS45 based Juju models
are deprecated. Drop their addresses (server_argyle and server_ps45
respectivly). Change the default address to server_ps5 as any new
virtual-hosts will be created there. Finally, drop any explicit
mappings for server_ps5.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-07-08 11:10:56 UTC

swm: wft/promote-to-proposed -- block promotion on the signing-signoff task

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-03-31 15:02:25 UTC

cranky: crl/git -- ensure that the git upstream pointer is correct

When we clone/checkout/reset a branch to update it ensure we set the
upstream pointer to the upstream remote/branch so that we can push the
branch back to whence it came.

At the same time configure the local repository to load cranky.config
when you are on a cranky/* branch. That config is pre-loaded with
configuration to default `git push` to `upstream`.

Then a simple `git push` on a checked out `cranky/*` branch will push to
the branch from which `cranky checkout` pulled it.

Protect all of this under a feature flag 'checkout-set-upstream'.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-07-08 12:31:15 UTC

kernel-stable-board: add cycle limits

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-05-20 21:24:02 UTC

kernel-stable-dashboard: elide all -s messages from dashboard

We have been protecting the -s option in case it affected the rest of
the board. This should be ok, make the option global.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-03-16 12:35:13 UTC

swm: snap -- request manual builds (disabled)

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-08-12 13:03:28 UTC

swm: bug,manager -- add the concept of being unable to handle a tasks

When we cannot handle a task because some critical data is missing we
want to continue to process the rest of the tracker and the remaining
trackers. Add a new WorkflowBugTaskError exception to represent this.
When thrown expose it in the reasons.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-07-07 13:04:27 UTC

admin: admin-lp-git-permissions -- add nvidia-kernel-committers when require

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-04-19 09:48:03 UTC

swm: regression-testing -- track in PPA when early testing was requested

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-05-22 21:13:19 UTC

swm: testing -- sharpen and reneable missing kernel checks

Add a new iterator to look for packages in a pocket or later. Use this
to sharpen the missing checks. Renable them with this softer protection
in place.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-02-10 11:05:42 UTC

ktl/launchpad_cache: fix me instantiation

Fix the instantiation of me so it actually is a real link to me. This
fixes the previous attempt to work-around the "limitation" (compliance
with a non-intuitive standards definition) on redirects exposed in
python3 requests infrastructure.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-05-26 15:11:38 UTC

swm: new-review -- handle preapproval via Triaged

When we wish to preapprove the new-review (at the same time as
sru-review) we have been moving new-review directly to Fix Released.
However when the new-review is then scanned it immediatly is recinded
back to New because we have no cached prepare-id. Later it will move to
Confirmed in the normal way; leading to confusion.

Instead have the pre-approver move the new-review to Triaged to indicate
this form of approval. When we find it in this state move directly to
Fix Released, and promote the sru-review clamp to new-review; this is
the version they reviewed so it make sense to share this clamp.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-05-26 22:12:11 UTC

swm: *-review -- recind update state only when changed

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-12-08 11:41:42 UTC

swm: tbt -- use the new --error-ok escape to return true failures

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-08-04 09:38:45 UTC

swm: wft/promote_to_proposed -- poll waiting for pockets to clear

When we find a kernel in our destination pocket request periodic rescan
so we detect when it becomes clear.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-08-12 20:25:42 UTC

swm: package -- handle null publishing date

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-11-26 13:51:07 UTC

swm: bug -- record the jira issue indicated via tags

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-12-08 00:23:15 UTC

dashboard: dashboard-update-html -- python3 porting

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-01-12 20:32:17 UTC

swm: package -- instantiate pkgs on first use

We are wanting to use this value ealier in the game. To simplify
initialisation dependancies switch to instantiating this on first use.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-03-30 21:44:44 UTC

swm: package -- convert debs version-clamps to general clamp support

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-03-23 11:23:37 UTC

swm: wft/workflow -- generalise hold-promote-to-updates blocks cascade

Allow us to indicate that a source should actually block its parent
kernel in a similar manner to that in which a snap tracker blocks its
parental tracker.

Add a new kernel-series swm tag block-parent-release which allows us
to note that this is such a source.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-11-08 21:22:08 UTC

swm: wft/prepare-packages -- PPA blockage is less interesting than in cycle blockage

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-11-19 14:58:27 UTC

swm: source-blocks -- ensure mirror flags are updated in all cases

We care very much that the mirror flags are updated to match the current
state when we save the tracker. External processing is relying on those
being accurate to optimise handling. Ensure we write them back on every
status change regardless of whether we scanned against those tags. Note
that if the flag is unchanged then swm-properties remain unchanged and
their save will be avoided.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-10-08 13:23:57 UTC

swm: tbt -- supply base ctx connections

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-03-22 20:18:54 UTC

ancillary-bot: limit lookups to those packages which express ancillaries

Use kernel-series.yaml to tell us whether a specific source has an
ancillary (generate for signed, or lrg, lrs for lrm). If not then we
can simply ignore the package.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-07-04 08:29:53 UTC

adt-matrix: MISS latest handling is a black art so disable

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-04-29 10:29:07 UTC

proceedures: series-opening -- snap store relates to LTS+1 only

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-12-13 21:56:32 UTC

stable/dashboard: do-sru-dashboard -- support external publishing

Support publishing the resultant dashboard to an external location.
Parameterise for the location. Generate the resultant dashboard payload
into that location and link the support files into the same location.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-01-24 23:05:24 UTC

swm: manager -- switch to status.json as primary data source

Switch to reading and writing the status.json as our primary data
source. The status.yaml becomes a legacy copy. We will remove the
yaml data once sufficient time has passed.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-12-17 21:36:37 UTC

swm-jira-sync: mark Cycle Status kernel-status

Ensure we tag our status entries as soon as possible. This helps us
identify and colourise them in the board.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2022-01-13 11:50:38 UTC

proceedures: README -- update for Jira

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-12-13 14:15:45 UTC

latest-tarballs: drop hwe kernels for wireguard-ci

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-11-30 14:39:17 UTC

mainline-builds: cod-ubuntu-config -- switch to using the internal version number

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-12-16 14:55:46 UTC

swm-ls: sort out owner/assignee lengths

Use the existing measurement support for calculate lengths for these
fields.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-08-09 08:49:25 UTC

swm: snap -- fix various channel-map caching issues

We have a number of issues with the caching of the snap-store data for
the snaps.The caching in `channel_revision` is just broken, but
luckily we always call `channel_version` first which does it correctly
and causes the former to function correctly.

Separatly we are not caching the channel-map for the last-published use
case. This means that the scan we do to determine our snap tracker
disposition could be out of sync with the scan we record for monitoring
purposes. This creates a race window in which the disposition could
change without us noticing.

Fix all of this by directly caching the channel-map at the `channel_map`
level.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-12-09 14:11:42 UTC

announcer: announcer-key -- formalise name of the key mapper

The current announcer binary is actually the key mapper for the
announcer and responds to announcer.key similarly to any of the other
destinations. Sort out its naming to make it consistent with
the others to simplify deployment.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-08-05 11:59:02 UTC

swm: wft/verification_testing -- stop monitoring verification-testing once in proposed

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-05-14 11:08:44 UTC

swm: wft/syn_prepare_packages -- make true failures more obvious

Use the Stalled meta-status to make failures in build more obvious on
the dashboard et al.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-07-25 23:06:50 UTC

swm: swm-publishing -- FilterSwmStatus use factory.ss

Share the common SwmStatus object as instantiated by the factory.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-07-06 09:48:16 UTC

swm: wft/syn_prepare_packages -- always use proposed now signing is cleaned

canonical-signing-bot keeps the signing PPAs clear to allow signing of
older versions to complete without version constraints firing, and to
keep the size of these PPAs from spiraling out of control. This removal
affects the nieve copy delta calculation. Always use proposed as our
destination.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-05-13 07:47:16 UTC

swm: package -- built needs to be filled in when present

We need to record a build location as soon as we find the published
source. We need that information for sru-review which is valid before
the builds complete.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-06-02 08:38:12 UTC

swm: package -- detect logless failures and retry them

Builds in `Failed to build` which are also logless should be
retries unconditionally.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-04-20 13:54:33 UTC

swm: wft/promote_to_proposed -- wait for promote-to-proposed when signing

We are waiting always for :prepare-packages to complete but this is
only correct for promote-to-proposed. For promote-signing-to-proposed
we need to wait for promote-to-proposed to complete. Move this to the
object config.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-07-06 08:08:17 UTC

swm-ls: add --sort-phase to select phase ordering

List tasks in phase order, including entry and exit from phases.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-11-17 21:51:22 UTC

swm-jira-sync: add run script to handle proxies et al

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-09-27 19:19:51 UTC

kernel-stable-dashboard: include snap testing tasks in selectors

Include snap-certification-testing and snap-qa-testing in the selectors.
Drop the snap- prefix so we tend to merge when sensible.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-04-20 08:42:14 UTC

swm: package,wft/promote_to_* -- switch to monitoring the delta

Determine the packages which are in a promotion and only monitor those
across the lifetime of the promotion. Record this information in the
tracker in the delta element, separated by promotion type. This allows
review tooling to present only packages which are in flight.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-09-03 10:25:28 UTC

swm: swm-instantiate -- look up ubuntu as a distribution

We are using the ubuntu object as if it is a distribution but looking it
up in projects. Look it up in the right namespace. Failing to do this
causes getSeries to use the wrong cache accessor.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-11-19 12:17:17 UTC

cranky: update-dkms-versions -- allow mappings to contain multiple components

Allow a mapping file to list more than one content fragment. The
individual elements will be concatenated locally.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-11-19 12:17:17 UTC

cranky: update-dkms-versions -- allow mappings to contain multiple components

Allow a mapping file to list more than one content fragment. The
individual elements will be concatenated locally.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-11-05 10:12:36 UTC

create-{sru-cards,respin-card,kernel-tasks}: install Jira versions

Switch over to Jira versions of create-sru-card, create-respin-card, and
create-kernel-tasks.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-10-04 20:19:55 UTC

swm: bug -- use kernel-series owner when present

When we have a kernel-series owner use that as the owner of a tracker.
Selection algorithm:

1) if workflow task has assignee use that, else
2) if kernel-series source object has assignee use that, else
3) if prepare-package has an assignee use that, else
4) no owner.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-07-06 16:31:05 UTC

swm: wft/workflow -- separate snap-prepare phase

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-09-14 11:18:13 UTC

ktl/bug_spam: add the triggering package and version to spam messaging

When we request verification in a series it is helpful if we also
include the package name and version we are spamming for to allow people
to decide what if any testing they intend against that specific package.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-06-17 12:29:23 UTC

snap: handle resource-not-found errors as a store error

Handle resource-not-found error returns as a store error. We have no
information about the snap so we cannot determine its disposition. It
is best to fail completely than miss-categorise it.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-06-30 22:24:43 UTC

swm: package -- handle archive lookup failures in debug output

When we fail to lookup and archive we end up with a null archive
pointer. Attempting to print the reference will crater. Detect and
handle this.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-07-09 13:03:30 UTC

web: Makefile -- install result files atomically

Install various output html files atomically at the end of the run.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-08-31 10:14:53 UTC

swm: swm-trello-sync -- implement sru-review rejected

When we perform an sru-review and the kernel is rejected we now set the
sru-review to Invalid. When this happens we want the trello card to
move with it to Rejected. Add support for tracking this task.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-09-06 14:00:09 UTC

swm: tbt -- add support for --ppa and --op

We would like to be able to request full testing against the early
builds for the main kernels in the PPA. Add `--ppa` and `--op <op>`
to allow explicit selection. Default `--ppa` to `op=boot`.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-09-03 20:23:44 UTC

swm: manager -- document the causes of PreconditionFailed exceptions

Make it clear what might cause use to see a PreconditionFailed exception
when processing a tracker.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-09-02 20:07:54 UTC

swm: manager -- drop long redundant migration code

Many moons ago we migrated all of the timestamps into a manager subkey
to simplify copying those forward into the new tracker status on update.
This was needed until all trackers with the old form were converted, ie
about 4 hours. Finally drop this.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-09-02 13:02:43 UTC

cranky: docs/cranking-the-kernel -- document the update-dkms-versions

Update the cranking documentation to include the new
`update-dkms-versions` script and move it to the appropriate
position in the flow; it must follow `cranky link-tb`.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-06-16 13:11:21 UTC

swm: wft/workflow -- limit snap consistency checks to published risks

We currently report snaps as inconsistent if any risk level is out of
step. This makes no sense if we have not yet released to that risk.
Only check risks we believe we have published to for consistent versions
of the snap.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-07-05 10:47:31 UTC

swm: manager -- fix unknown workflow task formatting

When we hit an unknown task in a tracker we generate a notification
"reason" so that it gets reported. This is incorrectly formatted
leading to confusing output in the dashboard. Reformat to include the
gross-state and flags field.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-04-21 23:11:12 UTC

swm: wft/prepare_package -- allow prepare-package-lrm to indicate master prepared

When waiting for a master bug to show prepared we need to consider the
likely top level components we are waiting for. These are the main
package and lrm. Monitor both.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-04-19 21:24:44 UTC

swm: package -- pocket_route avoid ancillary routes

We use pocket_route to work out which archive in a route we are using.
This is used when emitting testing requests. Those requests only make
sense against the PPAs in which the main kernels are built. Exclude the
ancillary routes in our search.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-06-25 19:45:17 UTC

swm: wft/workflow -- include won't fix in terminal states for testing phase

We are treating Won't Fix as Skipped, so this is a valid terminal state
for testing. Include it so we leave phase Testing.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-07-01 12:14:48 UTC

swm: wft/kernel_snap -- raise tag errors as crank errors

If we get a GitTagError while trying to enumerate our tags we no longer
know if we have a tag or not. Rather than getting into a mess throw the
tracker into CrankError.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-06-23 10:51:47 UTC

admin: admin-lp-sync-kernel-packages -- new helper to keep kernel-packages in sync

Sync the contents of kernel-series into launchpad kernel-packages team
subscriptions in line with the MIR process.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-05-09 06:55:27 UTC

swm: snap -- ESM snaps are on canonical-kernel-esm

Launchpad limits where a snap object can be if it is going to access
private objects; it must be on the same team as those private objects.
With uc16 now in ESM its kernels have moved to the ESM project. Expect
the snaps for kernels in ESM to be there too.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-05-11 14:26:25 UTC

swm: wft/promote_to_proposed -- handle cycle_hold in new

We are bouncing back and forth between New and Confirmed when cycle is
marked on hold. Detect the hold in New and remain there.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-04-07 15:09:16 UTC

swm-publishing: MonitorLaunchpadArchive -- skip on unknown archive reference

We will crater later in the game if we try and handle an invalid archive
reference. Detect this early so we will skip this archive; there can be
no change to detect in this archive.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-04-01 20:34:45 UTC

swm: wft/promote_to_proposed -- ignore component checks when no main package

For partial promotions we have no main package to check against. The
disposition of the main package determines the disposition for the
dependant packages. For now elide the check.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-08-09 17:17:39 UTC

kernel-stable-board: reorder generation to place style in the header

Style elements should be inside the document. Pull them down and into
the <head> element.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-08-05 07:43:22 UTC

kernel-stable-dashboard: link rt: and bt: directly to the series/package result

The testing dashboard now has ids on the rows to allow direct jump to
results; these are formed <series>:<package>:<version>:<op>. Form an
appropriate fragment (#...) to point to these.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-02-08 12:15:09 UTC

swm: packate,wft/promote_to_* -- track packages to promote

When we move a promotion task to 'Confirmed' determine which packages
are missing from the destination. Record that list as a promotion hint
in "delta" indexed by the taskname. Recording them by taskname allows
us to better follow the history of the promotions and in principle lets
us have multiple such operations in play at once (such as might happen
with promote-to-proposed and promote-signing-to-proposed with an
embargoed package such as lrg).

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-02-12 14:24:34 UTC

swm: wft/prepare_package -- detect early close of prepare-package-*

If a prepare-package task is closed early manually or (more commonly) if
the package is replaced in the build route without informing swm (by
updating the related version) the prepare-package can not be considered
complete. Start monitoring the builds when in Fix Released and should
they no-longer be present pull the task back to In Progress. This will
trigger missing tag and missing upload reports in the dashboard as
appropriate.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-07-16 08:49:27 UTC

swm: wft/verification_testing -- add verification-testing-{passed,failed} overrides

For all automatically updated testing tasks we should be honouring
consistent override flags. Add verification-testing-{passed,failed}.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-08-02 20:12:42 UTC

swm: bug -- catch exceptions when loading swm-properties

If the bug swm-properties field fails to load we can end up losing the
data contained in it. This leads the bug to become a "combo" tracker
and much of the key associated data is lost. Detect these failures and
convert them into WorkflowCorruptErrors. As we could not read the
swm-properties we cannot sync this into the bug without further damaging
the swm-properties, so we crowbar a crank-failure into the reasons field
in SwmInfo so that it will at least show up on the dashboard.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-03-19 08:27:16 UTC

swm: swm-publishing -- summarise current queue status

We currently report new entries arriving on the queue. This however
does not give you a feeling for how many remain to be cleared. Detect
all of the pending queue entries for each queue and produce a delta
relative to empty in each announcement, only annoucing on change.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-02-28 07:26:39 UTC

swm: package -- if source is invalid we cannot check test flavours

The data needed to generate the meta mapping is stored in kernel-series.
If we have no valid kernel-series source pointer we will crater trying
to generate it. Return a null list. We will catch the bad source
linkage elsewhere and put the whole tracker into a crank-failure so it
will never be consumed in anger.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-02-02 09:29:37 UTC

swm-maintenance: elide positive records by default

Now that we are showing postive records (currently building etc) we need
to hide positive records by default. Also check the actual status not
the reported one when checking. Use --all to include them.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-02-26 13:23:23 UTC

swm: bug,package -- separatly track the version of the source

It is important that we revisit the cached package versions if the
primary version on the tracker changes. Currently we check that against
the main package version as those are they same. However this breaks
down if we do a partial respin, such as a meta-only or LRM-only respin,
as we may not have a main package. Add a new `source` entry to the
versions table which tracks the title version and use it for title
validation.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-03-23 02:17:44 UTC

swm-publishing: instantiate our PPA access key

The act of giving PPA access to a user (such as the swm bot) does not
actually allocate a key. The key is only instantiated on the first
viewing of the subscription. This occurs when we view the subscription
on a personal +archivesubscriptions page via the
getArchiveSubscriptionURL() call.

Note that we must work around a python3 launchpad lib redirect handling
issue when using the lp.me link. The HTTP spec says that any redirect
must be followed using GET which breaks many of the API calls when
applied to the lp.me which is a redirect to the logged in user. This
link must be manually followed to avoid this.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-07-30 08:16:28 UTC

kernel-stable-dashboard: add tooltip for sr:

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-07-30 02:48:59 UTC

swm: wft/prepare_package,wft/workflow -- report derivative blocks only when active

When a tracker is marked as kernel-block-derivatives only actually
report this for the tracker when a derivative is actually blocked by it.
Use interlocks to detect this.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

Author: Andy Whitcroft
Author Date: 2021-02-22 14:05:16 UTC

swm: wft/promote_to_updates -- pull back if signoffs recinded

If a signoff is no longer verified then we should pull back any
unstarted promote-to-updates task.

Signed-off-by: Andy Whitcroft <apw@canonical.com>