~canonical-kernel/+git/kteam-tools:roxanan/cranky-close-testing-tweaks

Last commit made on 2023-12-12
Get this branch:
git clone -b roxanan/cranky-close-testing-tweaks https://git.launchpad.net/~canonical-kernel/+git/kteam-tools
Members of Canonical Kernel can upload to this branch. Log in for directions.

Branch merges

Branch information

Name:
roxanan/cranky-close-testing-tweaks
Repository:
lp:~canonical-kernel/+git/kteam-tools

Recent commits

3baf0f7... by Roxana Nicolescu

cranky/fixes.d: Temporary delete all the other fixes

Signed-off-by: Roxana Nicolescu <email address hidden>

0b928a4... by Roxana Nicolescu

cranky/crl/git.py: Include private kernels manually so the we can fetch them

For older cycles, kernel-series.yaml does not say that these kernels are
private.

Signed-off-by: Roxana Nicolescu <email address hidden>

64964fc... by Roxana Nicolescu

cranky: close: Fix base_version calculation for respins in 4/2 cycle

Due to 4/2 cycle, in case we need to respin both security main kernel
and sru main kernel, we end up with interleaving versions:
security spin-1: 6.2.0-11.11
sru spin-1: 6.2.0-12.12
security spin-2: 6.2.0-13-13
sru spin-2: 6.2.0-14.14

When cranky close is run for a derivative in the security cycle, both
spin-1 and spin-2 changes are there, but for the sru cycle, only spin-2
changelog (6.2.0-14.14 in this case) are there, instead of both spin-1
and spin-2. This happens because we look for the latest parent tag present in
the changelog (in this case it's the security spin-2: 6.2.0-13.13) and
the changelog is calculated from there.

This commit addresses this issues, and should work correctly for cases
like this, but also normal case where just a spin-1 is present.
We find the latest derivative tag and we check the common ascensor
between HEAD and this tag which should result in the master version of
sru-spin-1.

This only works for direct directives from master, therefore the old
implementation is left for second level derivatives.

Note: The latest tag is calculated based on changelog, not by using `git
tag` command, so that we could test it easily
with previous versions and it would not interfere with newer tags
already created. For backports, '~' is replace with '_'.

Signed-off-by: Roxana Nicolescu <email address hidden>

3bbd2fc... by Andy Whitcroft

info/kernel-series: Claim some Google Cloud kernels

Signed-off-by: John Cabaj <email address hidden>
Signed-off-by: Andy Whitcroft <email address hidden>

5607eda... by John Cabaj

info/kernel-series: Claim some Google Cloud kernels

Signed-off-by: John Cabaj <email address hidden>

02576a1... by Andy Whitcroft

cranky: promote-snap -- add support for promoting by tracker

Add support for promotion by tracker number. As things stand
you have no way to perform the promotions wanted for a stream#1 tracker
and not a stream#2 tracker. Add support for promoting by tracker to
simplify selection.

Signed-off-by: Andy Whitcroft <email address hidden>
Acked-by: Stefan Bader <email address hidden>
Signed-off-by: Andy Whitcroft <email address hidden>

7d0d432... by Ian May

info/kernel-series: Transition nvidia-hwe-22.04-edge from nvidia-6.2 to nvidia-6.5

Signed-off-by: Ian May <email address hidden>
Acked-by: Jacob Martin <email address hidden>
Signed-off-by: Ian May <email address hidden>

10d9b87... by Ian May

info/kernel-series: Transition nvidia-hwe-22.04-edge from nvidia-6.2 to nvidia-6.5

Signed-off-by: Ian May <email address hidden>

c1ce026... by Andrea Righi

annotations: update: properly update note for arch-flavour configs

Values in the form "arch-flavour" are incorrectly processed in the logic
to determine if the note needs to be updated, causing a bunch of notes
to be incorrectly replaced with 'TODO: update not'.

Fix this by searching the values with the proper arch and flavour.

== Additional notes ==

Inside the affected function we are replacing flavour with "arch-flavour" to easily access the items in the dictionary `self.config[conf]['policy']`.

But we also used the replaced flavour variable to search for the old value (and trigger the "update note" condition), that ended up being replaced another time, producing "arch-flavour-flavour", that is incorrect.

This is why all the notes for the values in the form "arch-flavour" were incorrectly replaced with the "TODO", because it was always failing to find the previous value.

Signed-off-by: Andrea Righi <email address hidden>
Acked-by: Stefan Bader <email address hidden>
Signed-off-by: Andrea Righi <email address hidden>

180b58b... by Andrea Righi

annotations: update: properly update note for arch-flavour configs

Values in the form "arch-flavour" are incorrectly processed in the logic
to determine if the note needs to be updated, causing a bunch of notes
to be incorrectly replaced with 'TODO: update not'.

Fix this by searching the values with the proper arch and flavour.

Signed-off-by: Andrea Righi <email address hidden>