-
Notifications
You must be signed in to change notification settings - Fork 154
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Modernize and fix with recent versions of PouchDB. #274
Conversation
…rotections for them like in CouchDB). No regressions, but JS test coverage isn't complete yet.
…ir released version.
I just restarted the failing tests for the third time to be sure, but I think the failures might actually be genuine. |
Reordered package.json so that packages that belong (or should belong) to the same monorepo are adjacent. Hopefully this makes it more obvious that they should be upgraded in lockstep.
Sorry @marten-de-vries, I had a green build at the time of the PR, but I then realized a few things:
That meant that plugins that live in the monorepo were properly tested against 6.4 since they require the Now I've pinned everything to 6.4, including the test suite, but since I've done that I've been going deeper and deeper down the rabbit-hole. Hopefully I've just managed to fix the sqlite build. Now I've got to figure out what's going on with |
Nice work. I looked into why
Not sure what the easiest way to actually fix this is. |
I'm really puzzled by those First things start to go sour after a timeout, but since the timeout is always reproducible and doesn't seem time dependent (I tried to vary the timeout duration) I guess it's due to a deadlock. Second, the test suite can't recover from that failure: I modify I commented this failing Commenting the whole
|
Regarding the CouchDB At least now we have an actual failure:
|
If the If we want to preserve backwards compatibility we could only change the test instance of At least we are seeing the light at the end of the tunnel, we just need to turn my last commit into the preferred option, wdyt @marten-de-vries? |
The following works:
I'm not sure about |
I think the proper fix would be to change isCouchDB to something that instead detects what it is used for: whether validation is supported. Renaming the function would be a nice bonus. We should probably open an issue for that on the pouchdb repo. But as that would take at least another release to be usable, I'm fine with whatever workaround you come up with for now. Including enabling a few extra includes. |
CI doesn't seem to pick up changes. |
The description of
I went with the hybrid approach, I changed the |
Actually, Anyway, we can fix that later. I'm going to merge this. |
Thanks for all your hard work on this, @loic! |
@marten-de-vries thanks for the merge! Seems like I jumped to conclusions too quickly re the replication stuff. I'm gonna make an additional 3 PRs:
|
That sounds like a great solution. |
No description provided.