Fabio, If you're doing it for the sake of the relations between Castle and NHibernate, you're doing it for the wrong reasons. The ILogger abstraction is so obvious that NHibernate stick with it. The other option (ILog) is normally used by the logging frameworks itself... surprises me that you didn't realized it.
But if you really thinks that an ORM should leak his internal logging stuff all over place, demanding it users to change their classes, you're a more limited leader than I thought. PS: A youtube song? Really? Even for you, it's was ridiculous. Cheers, Henry Conceição On Mon, Sep 20, 2010 at 3:35 PM, Fabio Maulo <[email protected]> wrote: > 2010/9/20 Henry Conceição <[email protected]> >> >> The issue isn't related with Windsor. The issue is that the ILogger >> abstraction is the most obvious one, and any >> project/application/framework that tries to decouple/abstract it's >> logging machinery ends up using it. > > You are talking about "any project/application/framework" you have saw. > I'm talking about "any project/application" I have worked and about "any > framework" you have the example of a framework whose only target is the > logging abstraction and it does not have neither the ILogger nor > ILoggerFactory interfaces. > btw, please, avoid to insist. You have what you asked and the only one > reason for this modification is the good relation between NH team and Castle > team (even if you have not realized it yet). > -- > Fabio Maulo > >
