Add missing got_static_announcement

[david415]

i copy pasted this snippit of code from my dev branch introless-multiintro_yaml_config.2
it's probably wrong... and needs tests; DO NOT MERGE

[coveralls]

Coverage decreased (-0.04%) to 92.231% when pulling 8f04b9a on david415:2801.fix_connections_yaml.0 into 6f8c96e on tahoe-lafs:master.

[coveralls]

Coverage decreased (-0.003%) to 92.16% when pulling 327d70f on david415:2801.fix_connections_yaml.0 into 85cf1d6 on tahoe-lafs:master.

[warner]

key_s should always be present now (since we got rid of v1 introducers). Let's remove the if and do these precondition checks all the time. Let's also make the assert on the service-name into a precondition(). (we can make this change to _got_announcement() too)

[warner]

I kind of liked the approach we had before, where NativeStorageServer created it's own Tub internally. Why move Tub creation up to the StorageFarmBroker? (in the future, we'll add HTTPStorageServer, which won't need Tubs at all, so it seems cleaner to do the Tub creation inside the class that knows it needs one). I'll have to think about this.. I can see that there's a question of what happens when a server is updated (e.g. we get an announcement for a server that we've already started using, so there's already a Tub for it: having the StorageFarmBroker own the Tub would make it easier to keep using the same Tub even though the NativeStorageServer gets replaced).

Is there any way we can merge got_static_announcement with the existing _got_announcement? There's a lot of code duplication between them, and I think merging them would teach us how to get the ownership/updating part right. got_static_announcement is lacking the s.on_status_changed hook that helps Magic Folders know when to start, the if old clause shouldn't ever fire for static announcements because they're added before anything else has been added. Except for the self.static_servers tracking, I don't think I see anything that would prevent got_static_announcement from just immediately delegating to self._got_announcement.

Please rebase to current master when you get a chance, this overlaps a couple of moderate-sized branches from the last month.

[coveralls]

Coverage decreased (-0.02%) to 92.148% when pulling 1f8e4c8 on david415:2801.fix_connections_yaml.0 into 61eb839 on tahoe-lafs:master.

[david415]

ok i've made the corrections you specified and removed code duplication. please review

[warner]

Some more items:

• the new diff removed the old.disownServiceParent() from the replace-an-old-server case, and removed the s.setServiceParent() from the new-server case. I think both need to be reinstated. Do a git diff master..HEAD and make sure that nothing's been changed that wasn't necessary. The name change from serverid to server_id could be put into a separate patch, but I think it's ok to have it combined into the rest of the patch too, no big deal.
• please rebase to current master and force-push to this branch.
• consider squashing all of this. in fact, maybe you should just extract a diff and then make a new branch that applies the entire diff as a single patch:

git checkout 2801.fix_connections_yaml.0
git diff master..HEAD >p
git branch new master
git checkout new
patch -p1 <p
git add, git commit
git diff new..2801.fix_connections_yaml.0 # this should be empty
git checkout 2801.fix_connections_yaml.0
git reset --hard new
git push -f origin 2801.fix_connections_yaml.0

That might be easier than rebasing this, and will result in the same code.

[coveralls]

Coverage decreased (-0.0008%) to 92.163% when pulling 78e77c3 on david415:2801.fix_connections_yaml.0 into 61eb839 on tahoe-lafs:master.

[warner]

looks good.. just waiting for tests to complete, then I'll land it. thanks!

[coveralls]

Coverage decreased (-0.008%) to 92.156% when pulling de61cd2 on david415:2801.fix_connections_yaml.0 into 61eb839 on tahoe-lafs:master.