rbt

This is trickier than it first appears, particularly since nearly everything we do with Pg is done via a prepared statement and replacing ? or $1 with the DEFAULT keyword...

The DBIish tests try every database it knows about but they're supposed to fail gracefully when a connection cannot be made for any reason (missing libraries, network connectivity, username/password failure,...

That's quite helpful. So it's actually an SQLite threading failure. I have neither a Mac or SQLite expertise, so I'm not much help there. But since you're intending to use...

The thread test uses a single database connection across several Raku threads, which fails if locking isn't properly done.

I've pushed a few improvements, with help from abraxxa. I think it documents everything that is fully fleshed out and not marked deprecated. If you have additional suggestions I'm happy...

Right now everything goes through PQexecPrepared which is limited to CRUD type statements. We should have a fall-back mode for Pg which does parameter quoting/injecting and calls PQexec.

dbh.Converter is sufficient for extracting data from the database but there needs to be another converter which runs during sth.execute() for importing data. Since these will probably use dbh assistance...

There have been some improvements on this over the past year, mostly (I think) from the new dispatch mechanism being used by NativeCalls. With PostgreSQL using your test code (driver/username/password...

A quick look at the DBIish MySQL driver doesn't show anything SSL related so I presume it's something between libmysql and the backend. It seems there are a number of...

Ah. I see. There are even commented out regression tests from the perl driver which haven't been ported for these (mysql_init_command specifically). mysql.connect() should take an *%params arg similar to...