You use a third-party library/framework but don't explicitly specify in the code how you want logging to be done. Which would you prefer?
-
-
Replying to @propensive
a lib should provide logging effects that are tailored to its function. I'm not convinced one can abstract over this.
2 replies 0 retweets 1 like -
Replying to @tpolecat @propensive
Logging is not a concern clients should be burdened with, it does not belong into the API.
1 reply 0 retweets 0 likes -
-
Replying to @propensive @tpolecat
Caller–callee, client–service, what’s the big difference? It is all about modularization and about responsibility.
1 reply 0 retweets 0 likes -
Replying to @rolandkuhn @tpolecat
Wasn't sure if you meant clients to be just "users of library X"...
1 reply 0 retweets 0 likes -
Anyhow, do you not think it's a problem having logging configured in some uncertain place you can't debug like code?
1 reply 0 retweets 0 likes -
I'm imagining an implicit which could either handle logging itself, or delegate that to some other specific config.
1 reply 0 retweets 0 likes -
Replying to @propensive @tpolecat
I sympathise with the sentiment, but to me that is not principled. Logging is monitoring is an integration concern.
2 replies 0 retweets 0 likes -
Therefore it is validated by developers (debugging) and CI tests / CD ops.
1 reply 0 retweets 0 likes
But as a developer I'd still like to see the logs without doing any integration. A one-line implicit seems a good way.
-
-
Replying to @propensive @tpolecat
… then you cannot deploy those very same bits into production without the implicit becoming isomorphic to logback.xml
1 reply 0 retweets 1 like -
Replying to @rolandkuhn @tpolecat
But difference is that logback.xml is not a known authority on conf. Classpath order and and JVM params can obfuscate.
2 replies 0 retweets 1 like - Show replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.