OD-1593/queue/sequence groups

[pankk]

Do not merge, should be added to buildout once done.
Note: our buildout base for this repo is currently: fc7ea6a
I'll pull our 12.0 branch up to that version so reviewing will be easier.

For reference, the 12.0 branch was on dd05552 before.

[pankk]

Note that I did not add a dependency on this module because this won't break this module. Inheriting the Job class in a custom module for this did not seem reasonable for this kind of change. Albeit, this is a shim.

[pankk]

Here's an additional test I did which you can run in a shell. It covers testing automated invoice validation in combination with vanmoof_magento.
Note that it uses a cr.commit, so it will write to your database if you decide to run it.

from odoo.addons.sale_automatic_workflow_job.tests.test_auto_workflow_job import TestAutomaticWorkflowBase, TestAutoWorkflowJob
self = TestAutomaticWorkflowBase
self.env = env
override = {'validate_invoice': False}
workflow = self.create_full_automatic(self=self, override=override)
self.many_sales = [self.create_sale_order(self=self, workflow=workflow) for x in range(5)]
many_sales = self.many_sales[0].browse([r.id for r in self.many_sales])
values = {'street': 'street', 'city': 'city', 'zip': '1234zp', 'country_id': 1, 'phone': '21345613', 'email': 'email@vanmoof.com'}
[so.partner_shipping_id.write(values) for so in many_sales]
# [True, True, True, True, True]
self.progress(self=self)
invoices = self.env['account.invoice'].browse(
    many_sales.mapped('invoice_ids').ids
)
# Check related invoices are all draft
invoices.mapped('state')
# ['draft', 'draft', 'draft', 'draft', 'draft']
# Disable tier validation for now
self.env['tier.definition'].search([]).write({'active': False})
self.env.cr.commit()
# Run automatic workflow to validate invoices
[env["automatic.workflow.job"]._do_validate_invoice(invoice) for invoice in many_sales.mapped('invoice_ids')]
self.progress(self=self)
# We can mock the queue job instead if we don't want to commit to our database
self.env.cr.commit()
invoices.mapped('state')
# Wait for all invoices to get validated
invoices.refresh()
invoices.mapped('state')
# ['open', 'open', 'open', 'open', 'open']
# In the logs you should see entries such as:
# odoo.addons.queue_job.jobrunner.channels: job d28819d4-7492-4e4e-84cf-0598cdf8031e re-queued because a job with the same sequence rule 'invoice' is already running in channel root(C:2,Q:1,R:1,F:0)
# which indicates that the jobs with a matching sequence rule were deferred

[antonioburic]

just asking to get a better grasp on it, still in the process of understanding this flow - in this case the value of a different rule can be set then the rule name set above (first one taken), is this expected or it doesn't matter when we have multiple ones?

[pankk]

Only one rule is allowed per model:

queue/queue_job/models/queue_sequence_rule.py

Lines 37 to 41 in 3151ffa

	_sql_constraints = [
	(
	"uniq_model_id",
	"UNIQUE(model_id)",
	("Only one rule per model allowed"),

I left some stubs to allow for multiple rules per model, as I had started with that approach. I mid-way decided not to implement multiple rules per model because it would make the solution more complex than what we need, especially considering not making the job queue any slower.

[antonioburic]

tried to get as detailed look as I could, this will be a great way to manage it - really clean and efficient, nice work Kevin!