Clojurians Log v2
sql
I'm curious how many folks here use hikari-cp? And do you use it with clojure.java.jdbc or next.jdbc?
❎ I don't use hikari-cp
🟡 I use hikari-cp with clojure.java.jdbc
🟢 I use hikari-cp with next.jdbc
(I'll add one of each reaction to make it easier to vote)
I'm particularly interested in 🟢 in relation to https://github.com/seancorfield/next-jdbc/issues/312
My ❎ context is cljdoc, which uses SQLite, which (I think?) doesn't benefit from connection pooling.
I use every approach in different projects.
But any new code that I write does not use hikari-cp. And in other projects that I fully control, I gradually remove it when I get to it.
@p-himik Could you elaborate on why, and what you do about connection pooling in its absence?
• I prefer to avoid having extra dependencies if I can, even if they're small
• I use the next.jdbc's built-in support for HikariCP
I'm using HikariCP (not hikari-cp), with next.jdbc, for SQLite Connection Pooling (WAL-mode sqlite)...
• Reader DataSource is a Pool: number of connections calculated using HikariCP's formula
• Writer DataSource is a single connection.
Aside: Specific to SQLite, Anders Murphy's project is very promising: https://github.com/andersmurphy/sqlite4clj It does away with needing a JDBC driver, and provides pooling out of the box. Once it becomes beta / stable enough, I'll try to switch to it and subtract the JDBC deps from my apps.
Based on early responses, it sounds like it would be worthwhile providing some out-of-the-box integration with hikari-cp (in addition to the existing HikariCP support). Please keep the feedback coming!