Moisés López - https://www.vauxoo.com/ authored Aug 19, 2022 and Moisés López - https://www.vauxoo.com/ committed Aug 19, 2022

Newer Odoo versions is showing a WARNING when there are not tests to run

The log warning is the following:
 - `WARNING db odoo.tests.runner: 0 failed, 0 error(s) of 0 tests when loading database 'db'`

But new projects doens't have tests in the first time

So, the "warnings" pipeline could be raised

But it is not a warning easy to fix in the first step

So, it is better to ignore it

The new colorized log from tests using "unbuffer" needs to be parsed correctly

In order to detect errors, warnings and so on

In the past, the code was adapted to support colorized warnings

Now it is adapted to support all kind of colorized logs

Note: It was extracted from MQT and it was extracted from serverfault:
  - https://github.com/OCA/maintainer-quality-tools/blob/d705b9d68abd2/travis/test_server.py#L65
    - http://serverfault.com/questions/71285

Moisés López - https://www.vauxoo.com/ authored Aug 09, 2022 and Tulio Ruiz committed Aug 12, 2022

Allow to the user to know where the logs are saved
in order to check if next step is using the correct path

Moisés López - https://www.vauxoo.com/ authored Aug 08, 2022 and Moisés López - https://www.vauxoo.com/ committed Aug 11, 2022

Avoid raising error if the directoy was already created
Allowing to run 2 times the same command locally
The coverage content is created with file flag "w" instead of "a"
in order to be cleaned if it already exists
The coverage content was removed the extra spaces
The coverage is showed in the logs similar to old way and ran first

Compatible with docker remote where you do not have installed

Fixing the following error:
 - "ConnectionError: ('Connection aborted.', error(2, 'No such file or directory'))"

- Just in case someone use the main branch in the -dev group for any reason we detect if it is a dev repo so it don't push to latest, just to dev.

We do so by detecting if the branch name is the same as the version and now we also check if the group name ends with -dev, because we have as general rule to name the groups with 'group-dev' for development and 'group' for the main repo

- Expressions like  wheren't evaluated properly, the string was spllitting in the second = instead of the first.

If for some reason there is a file with the same name as the folder we are trying to create to store the image name the CI will show an error, now we need to remove such file so the mkdir can create the folder after.

[FIX] Pip is using p3 but default is p2

[FIX] ALLOW_DEPRECATED is not required as an env var

[FIX] Show deprecation count

[FIX] Remove debug message

[FIX] Do not count deprecation as general warnings

[IMP] Better log format and add debug information

Some modules need to install an extension so this is only allowed if the user is superuser. This won't work for production or testing envs.

If by any chance one of the possible hosts to scan is down or inaccessible we must warn the user instead of failing. If we continue there is a chance that the repo don't need any of the failing hosts, if it is needed will fail (which is expected at this point)

- We use stdout and stderr to avoid the double logging format
- Spurs detects better the errors during the build so is easier to handle in such cases

[FIX] Just in case check is there is any image

[FIX] Better use check_call because no output is shown for a while and we know what that mean

Some time the autoclean is excuted just before the CI have time to run the tests in which case an exception is shown, now we pull the image just before the postgresql container is created to avoid this.
A better message is shown just in case the postgresql image is now properly set.