Closed jcgsville closed 2 years ago
Huh! Looks like we weren't running prettier in CI?! Fixed in #138 - please rebase, then hopefully you can run the tests :)
@benjie I think this PR is ready to go with the exception of the release notes checkbox. Not sure if I'm missing where to add those?
What's changed since your 👀 were on this:
root: true
behave as intendedroot: true
behave as intendedApologies that I didn't realize force pushing to my branch would break Github's requested changes. The change you requested was a small tweak to the README which I implemented.
Finally, as I was working on the unit tests, I found a separate bug in makeRootDatabaseConnectionString()
. This should probably be a separate PR to fix this. If it makes more since for it to be one, let me know.
The 🐛 : when setting rootConnectionString to just the db name, the connection string used when root: true
incorrectly had a slash prepended to the database name.
For example, when rootConnectionString: "postgres", connectionString: "postgres://localhost/app_db"
the connection string when root: true
should be postgres
but it was /postgres
.
I noticed this because the unit tests default rootConnection string was just postgres
and the assertions were failing unexpectedly. I've modified settings.ts:420
to fix it, and I believe this change produces the desired behavior. But it's worth your 👀 to make sure I didn't misunderstand something. I also added a unit test to cover this case. I believe the existing unit tests on this function were sufficient to prove that this change introduced no regressions.
@benjie know of anything left before this can be merged?
I addressed the recommendations to fix the tests. Last thing pending is resolution surrounding the pathname vs dbname debacle
My new test didn't pass linting :(
I tried fixing it using prettier playground but that wasn't enough; please reformat :+1:
Thanks for your thorough reviews!
Of course it was the trailing comma :man_facepalming:
Released in 1.3.0 :raised_hands:
Description
Fix for issue 132. afterReset actions of type command now respect the
root
property in the action specPerformance impact
I can't imagine a performance cost to this
Security impact
None that I'm aware of
Checklist
yarn lint:fix
passes.yarn test
passes.RELEASE_NOTES.md
file (if one exists).I wasn't able to get my unit tests to run locally. Happy to try to write a test for this if you have any advice on what I'm missing? This was the behavior when I checked out the commit before mine so none of my changes are involved here.
Also, I don't see a RELEASE_NOTES.md, so I made no additions there.