From: Alexis Lothoré <alexis.lothore@...>

This series fixes regression report generation on "next" branches, as raised in
[1].

The first five patches are preparatory updates for the real fix, being either
refactoring, cleanup or unit tests addition to better understand how integration
branches are used in send-qa-email.
The proper fix is in 6th commit, followed by corresponding tests
Finally, the last commit add Alexandre's "next" branch as "fork" branches to
enable regression reports generation when testing patches, as suggested in [1]
too.

Since patch testing branches are force-pushed on a regular basis, it is quite
difficult to get a relevant testing scenario, so this series has been tested by
faking SHA1 in yocto_testresults_query to match some master-next results in
yocto-testresults at the time of testing this series. I would gladly take
feedback about this series running for real in a master-next branch

[1] https://lists.yoctoproject.org/g/yocto/message/59067

Alexis Lothoré (8):
scripts/utils: add unit tests for getcomparisonbranch
scripts/send-qa-email: remove unused variable
scripts/send-qa-email: invert boolean logic for release check
scripts/send-qa-email: protect is_release_version from None value
scripts/send-qa-email: add tests for is_release_version
scripts/send-qa-email: fix testing branches regression reporting
scripts/test_send_qa_email.py: add tests for base/target pair guessing
config: flag A. Belloni master-next branch as testing branch

config.json | 2 +-
scripts/send_qa_email.py | 34 +++++++----
scripts/test_send_qa_email.py | 31 ++++++++++
scripts/test_utils.py | 104 ++++++++++++++++++++++++++++++++++
4 files changed, 158 insertions(+), 13 deletions(-)
create mode 100755 scripts/test_utils.py

--
2.39.2

On Mon, 2023-03-13 at 15:51 +0100, Alexis Lothoré via
lists.yoctoproject.org wrote:

From: Alexis Lothoré <alexis.lothore@...>

This series fixes regression report generation on "next" branches, as raised in
[1].

The first five patches are preparatory updates for the real fix, being either
refactoring, cleanup or unit tests addition to better understand how integration
branches are used in send-qa-email.
The proper fix is in 6th commit, followed by corresponding tests
Finally, the last commit add Alexandre's "next" branch as "fork" branches to
enable regression reports generation when testing patches, as suggested in [1]
too.

Since patch testing branches are force-pushed on a regular basis, it is quite
difficult to get a relevant testing scenario, so this series has been tested by
faking SHA1 in yocto_testresults_query to match some master-next results in
yocto-testresults at the time of testing this series. I would gladly take
feedback about this series running for real in a master-next branch

[1] https://lists.yoctoproject.org/g/yocto/message/59067

Alexis Lothoré (8):
scripts/utils: add unit tests for getcomparisonbranch
scripts/send-qa-email: remove unused variable
scripts/send-qa-email: invert boolean logic for release check
scripts/send-qa-email: protect is_release_version from None value
scripts/send-qa-email: add tests for is_release_version
scripts/send-qa-email: fix testing branches regression reporting
scripts/test_send_qa_email.py: add tests for base/target pair guessing
config: flag A. Belloni master-next branch as testing branch

I think there is a regression somewhere:

https://autobuilder.yoctoproject.org/typhoon/#/builders/85/builds/2085/steps/29/logs/stdio

Cheers,

Richard

Hi Richard,
On 3/22/23 10:41, Richard Purdie wrote:

On Mon, 2023-03-13 at 15:51 +0100, Alexis Lothoré via
lists.yoctoproject.org wrote:

From: Alexis Lothoré <alexis.lothore@...>

This series fixes regression report generation on "next" branches, as raised in
[1].

The first five patches are preparatory updates for the real fix, being either
refactoring, cleanup or unit tests addition to better understand how integration
branches are used in send-qa-email.
The proper fix is in 6th commit, followed by corresponding tests
Finally, the last commit add Alexandre's "next" branch as "fork" branches to
enable regression reports generation when testing patches, as suggested in [1]
too.

Since patch testing branches are force-pushed on a regular basis, it is quite
difficult to get a relevant testing scenario, so this series has been tested by
faking SHA1 in yocto_testresults_query to match some master-next results in
yocto-testresults at the time of testing this series. I would gladly take
feedback about this series running for real in a master-next branch

[1] https://lists.yoctoproject.org/g/yocto/message/59067

Alexis Lothoré (8):
scripts/utils: add unit tests for getcomparisonbranch
scripts/send-qa-email: remove unused variable
scripts/send-qa-email: invert boolean logic for release check
scripts/send-qa-email: protect is_release_version from None value
scripts/send-qa-email: add tests for is_release_version
scripts/send-qa-email: fix testing branches regression reporting
scripts/test_send_qa_email.py: add tests for base/target pair guessing
config: flag A. Belloni master-next branch as testing branch

I think there is a regression somewhere:

https://autobuilder.yoctoproject.org/typhoon/#/builders/85/builds/2085/steps/29/logs/stdio

ACK, will take a look at it

Regards,
--
Alexis Lothoré, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com