Matthew Johnston
Matthew Johnston
@pavancse17 and @keatz55 right now if you keep your pool size at 1, this segfault won't occur. I'm still in the process of trying to figure out where the address...
I'm still looking in to this issue. Will need to probably utilize a custom compiled ERTS with valgrind compiled with it.
@ConnorRigby may have more insight into this than I do. I have not played much with the encrypted database stuff.
@ewildgoose I've actually looked at duckdb and am probably more inclined to make another package specifically for duckdb as opposed to trying to shim it in with exqlite. I too...
@mdwaud it seems like `exduckdb` _could_ be used for the `ecto_sqlite3` in place of this adapter. I tried not to tightly couple `exqlite` to `ecto_sqlite3` in case someone comes up...
I don't know what the performance implications of bumping that limit up is. We should go look at sqlite3 in ruby to see what it does for active record. I'd...
@kevinlang maybe we could figure out a way to support it just being left at default but also allowing someone who knows what they are doing to bump the limit...
@thomastay curious what were you doing to hit that limit? bulk inserting?
Interesting. Think you can mock up a dummy integration test for us and we can get this to be configurable / able to better handle that situation?
Quick question, does that same table definition work for postgres as well? What I think the issue is is that the unique index contains `[:device_id, "coalesce(sim_id, '')"]`, the unique index...