fix(db): db seed should try all known config file names by default

[btisdall]

Fixes #602

[mcollina]

Thanks for opening a PR! Can you please add a unit test?

[btisdall]

Oh sure. I considered it but thought it might be overkill – happy to though. EDIT: overkill on the basis that loadConfig already had unit tests, but actually it seems it doesn't.

[btisdall]

Hey @mcollina, want to make sure I'm setting off on the right track here. The current test of an acceptable config file name seems to rely on packages/db/fixtures/sqlite/platformatic.db.json. For a meaningful end to end'ish test one would need to try each of the supported files in turn, so I could see swapping this out for a file containing the source data and writing each out one in turn, with a teardown and/or temp directory for each one. The other way to go would be mocking, but I can see this is only used for undici usages rn and there's no general purpose mocking library in the project, which I'm going to guess is a deliberate decision. While I'm here, I'm a bit puzzled by the repetition of ourConfigFiles here, what's the reason for the hof supplying its own value? Thanks!

[mcollina]

Yes exactly, don't mock unless it's absolutely necessary.

While I'm here, I'm a bit puzzled by the repetition of ourConfigFiles here, what's the reason for the hof supplying its own value? Thanks!

Where is that repeated?

[btisdall]

It's first defined here.

[mcollina]

They are explicitly different, it's not repeated.

[btisdall]

Gaaah yes of course they are 🤦 . This might be related to the fact that I broke my glasses a week ago and haven't received the replacements yet 🙃 . Thanks!

[mcollina]

Can you please fix linting?

/home/runner/work/platformatic/platformatic/packages/db/lib/seed.mjs
Error:   1:10  error  'resolve' is defined but never used  no-unused-vars

✖ 1 problem
 ELIFECYCLE  Test failed. See above for more details.
Error: Process completed with exit code 1.

[mcollina]

You'd also need to fix DCO metadata (click on details for instructions)

[btisdall]

Thanks for your patience here @mcollina , between learning a new test framework, getting my head round a largeish project, the day job and the dog it's taking a while.

[btisdall]

RFR except I don't understand why this is still tripping the DCO check as I rewrote the first commit to include the signoff line 🤔 .

[mcollina]

I've checked all commits and they are all correctly signed off, not sure where the DCO app got stuck

[mcollina]

lgtm

[btisdall]

👀

[btisdall]

The test failure is in another package and I'm reasonably certain is unrelated to the change here (actually, I'm struggling to identify exactly which test has failed, I had to grep the logs for "not ok" and then the most helpful clue seems to be tapError: no plan.

[btisdall]

Thanks for re-triggering the tests. Unfortunately I don't have access to a Windows machine so I can't try to debug the continued failure.

[mcollina]

unfortunately we have a few flaky tests on windows