
Hi Logback team, yes it is already by design that there is No separation of main vs test logback configuration - http://logback.qos.ch/manual/groovy.html#ConditionalConfiguration But unfortunately this is a blocker feature for our team to switch to groovy from xml. XML model with logback-test.xml let developers be crazy with extending logs(appenders, custom loggers, custom ... ) for testing and be absolutely secure that it does not effect production mode. As changes in logback.xml file is kind red flag for detailed code review of code changes. Usage of special system/environment variables is not good too as it is additional argument for simple "mvn test", nobody wants additional complications for command line executions (manual, jenknis, .... ). please share a link where you describe a reason why "logback-test.groovy" is bad design. Can you reconsider your decision for "logback-test.groovy" ? thanks a lot, Roman Ivanov

Hi Roman, The problem of using different configuration files depending on context is an essential one. The absence of support for logback-test.groovy is not by design. However, I think that much of what can be accomplished by logback-text.* can be accomplished by judicious use of include files. Unfortunately, Gaffer (logback's groovy configurator) does not support include files either. I am planning major changes in the way Joran (the XML configurator) and Gaffer operate. Both configurators will work on a language independent model representation decoupling the configuration model and actions on that model paving the way for a whole range of new cool features, independent of language -- Ceki On 3/15/2016 18:18, Roman Ivanov wrote:
Hi Logback team,
yes it is already by design that there is No separation of main vs test logback configuration - http://logback.qos.ch/manual/groovy.html#ConditionalConfiguration
But unfortunately this is a blocker feature for our team to switch to groovy from xml. XML model with logback-test.xml let developers be crazy with extending logs(appenders, custom loggers, custom ... ) for testing and be absolutely secure that it does not effect production mode. As changes in logback.xml file is kind red flag for detailed code review of code changes.
Usage of special system/environment variables is not good too as it is additional argument for simple "mvn test", nobody wants additional complications for command line executions (manual, jenknis, .... ).
please share a link where you describe a reason why "logback-test.groovy" is bad design.
Can you reconsider your decision for "logback-test.groovy" ?
thanks a lot, Roman Ivanov
_______________________________________________ logback-dev mailing list logback-dev@qos.ch http://mailman.qos.ch/mailman/listinfo/logback-dev

Hi Roman, When I wanted more flexibility in the configuration of the logger, I wrote a class that read the configuration from any specified XML file, or even choose one of several different configurations from the same file based on machine host name or a JVM property. It's part of a library that has other stuff, and it's written in Scala but if anyone is interested, it isn't all that complicated. Donald Sent from my iPad
On Mar 15, 2016, at 13:18, Roman Ivanov <ivanov-jr@mail.ru> wrote:
Hi Logback team,
yes it is already by design that there is No separation of main vs test logback configuration - http://logback.qos.ch/manual/groovy.html#ConditionalConfiguration
But unfortunately this is a blocker feature for our team to switch to groovy from xml. XML model with logback-test.xml let developers be crazy with extending logs(appenders, custom loggers, custom ... ) for testing and be absolutely secure that it does not effect production mode. As changes in logback.xml file is kind red flag for detailed code review of code changes.
Usage of special system/environment variables is not good too as it is additional argument for simple "mvn test", nobody wants additional complications for command line executions (manual, jenknis, .... ).
please share a link where you describe a reason why "logback-test.groovy" is bad design.
Can you reconsider your decision for "logback-test.groovy" ?
thanks a lot, Roman Ivanov _______________________________________________ logback-dev mailing list logback-dev@qos.ch http://mailman.qos.ch/mailman/listinfo/logback-dev
participants (3)
-
Ceki Gulcu
-
Donald McLean
-
Roman Ivanov