@seancorfield in clojure.java.jdbc passing in a connection string to get-connection with mismatched parameters gives you a generic "Vector arg to map conj must be a pair". example "<mysql://localhost:3306/dbname?user=root&password=>" the person assumed this is what the connection string should look like if your password is blank. I was wondering if this is something worth wrapping in better error messaging or not.

Hmm, I would have expected that to work and just produce a nil password field. Can you open a JIRA issue? (I'm deep in debugging something unrelated right now)

will do

Include the stack trace -- it looks to me like the code should just defer to java.net.URI for that case...?

I'm running it at the repl using lein try and the stack trace is literally one line

Type *e to see the full stack trace

I've repro'd (since I had a REPL open and I'm stuck on my other problem! 😆 )

oh nice, I didn't see any calls to java.net.URI in the stacktrace

The bug is in parse-properties-uri which it calls after it gets the URI... I didn't write that function 🙂

Ah, (into {} query-parts) and that is produced by calling .split on foo= which doesn't produce a 2-element vector...

wow, you debugged it before I finished creating the JIRA ticket 👏:skin-tone-2:

I have a fix... testing...

OK, release 0.7.1 should be up on Maven later today with a fix for that. Thank you!

awesome! good luck on your other problem, hopefully getting an easy win with this one will help you get over the blocker