Fix Bolt support

[johannessen]

Resolves #20 (for real this time).

This doesn’t address #28, majensen/perlbolt#38, or johannessen/neo4j-driver-perl#11. Other than that, the test suite succeeds now on my system. And yes, this time I did double-check that it actually used Bolt.

---

Sorry that automatic merging isn’t possible. I’m afraid the CI workflow changes you made today prevent me from updating my fork. I keep getting a message saying refusing to allow an OAuth App to create or update workflow '.github/workflows/test.yml' without 'workflow' scope, but I’m not even using OAuth.

Could be a bug in GitHub, but I lack the time to investigate it further.

[majensen]

Hey Arne - I'm going to figure how to make the CI work for you (ideally for everyone); still in "beta" :-)

Really appreciate you putting the finishing touches on this.

[majensen]

@johannessen One thing to do -- go ahead and branch directly in the repo, since you're allowed

[majensen]

One thing - I want to leave in 0012_connect.t and its data files. This test mocks the server and the data files are fake responses. It was borking, but I fixed that in 331c68e. Can you pull and just patch ...Neo4j::Driver?

[johannessen]

Well, I made a bit of a mess with #27. Helping to clean it up is the least I can do.

About 0012_connect.t – the mocking is very nicely done, but after 9589d0e, it looked like a no-op to me. I may have misunderstood how that test works though, being less familiar with the Neo4p code than you are, obviously.

Anyway, I have now removed that particular commit and kept the other one.

[johannessen]

Hey Arne - I'm going to figure how to make the CI work for you (ideally for everyone); still in "beta" :-)

Right. Pushing to my fork succeeds again now – not really sure why. Maybe a config snafu on my end, or GH fixed something on theirs.

FYI, unlike Travis, the GH workflow also tries to run in forks, but fails, apparently because $DOCKERH is unset.

https://github.com/johannessen/rest-neo4p/runs/1808494080

[majensen]

Yeah that's tells me that actions won't use my envs on your fork, even if you're requesting a pull into my repo. Which I guess is ok, but this is a situation where I want my canned action, including its "secret" resources, to run on anybody's PR, which I would think is a reasonable and common use case. 🤷‍♂️

…

Sent from my iPhone

On Feb 1, 2021, at 12:22 PM, Arne Johannessen ***@***.***> wrote:  Hey Arne - I'm going to figure how to make the CI work for you (ideally for everyone); still in "beta" :-) Right. Pushing to my fork succeeds again now – not really sure why. Maybe a config snafu on my end, or GH fixed something on theirs. FYI, unlike Travis, the GH workflow also tries to run in forks, but fails, apparently because $DOCKERH is unset. https://github.com/johannessen/rest-neo4p/runs/1808494080 — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.