[WIP] Add rabbitmq msg for announcer

[dcermak]

This fixes #1834.

The whole code is untested, since I wasn't able to find any test code for announcer.py. I would be more than happy to add some tests to this, as I have absolutely no idea whether this works.

[coveralls]

Pull Request Test Coverage Report for Build 4553

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 28.183%

---

Totals
Change from base Build 4550:	0.0%
Covered Lines:	4184
Relevant Lines:	14846

---

💛 - Coveralls

[jberry-suse]

This would seem to be backwards. OBS should be emitting an event that the announcer listens for to send out an e-mail.

[dcermak]

In this case OBS is not sufficient, the factory-package-news script is invoked by openQA after a new changelog is created (I think). But I tend to agree with you, this process seems the wrong way around.

@DimStar77 you probably know more about this and what the right procedure is.

[coolo]

OBS sends out an event if openSUSE:Factory:ToTest/images is published - containing this ISO. I wonder what's your use case.

[dcermak]

I would like to run a build using the fresh Tumbleweed image, once it got released. Originally I wanted to listen for the rabbitmq event that ttm sends around, but Dominique told me that they are probably firing too soon and that OBS still has to do the release afterwards.

[DimStar77]

OBS sends out an event if openSUSE:Factory:ToTest/images is published - containing this ISO. I wonder what's your use case.

the problem at this time is that download.o.o will only be updated in like 2 - 3 hours, as we hold off the switch for mirrors to catch up. Hence the idea of the announcer also creating an event - as it actually monitors when the publishing is actually really done and ready for users to consuime download.o.o again

Preferably this event would come from another place, like the "publisher holding off"-script for example, at the time when it does the switch - but anybody daring to touch that code will curse for the rest of his life

IFF we would add rabitmq to the publisher scripts, then of course the announcer would/could/should listen as well

[jberry-suse]

/me remembers a proposal on this topic. oh well.

[coolo]

Preferably this event would come from another place, like the "publisher holding off"-script for example, at the time when it does the switch - but anybody daring to touch that code will curse for the rest of his life

now that explains @lnussel's behavior! :)

[dcermak]

Since I am quite inexperienced with the openSUSE infrastructure, what do you all think that should be done about this? If this is not the way forward, I'd rather close this and discuss a better solution in #1834.