inline On Sat, Mar 14, 2009 at 5:54 PM, Sidar Ok <[email protected]> wrote:
> Inline > > On Sat, Mar 14, 2009 at 12:58 PM, Ayende Rahien <[email protected]> wrote: > >> inline >> >> On Sat, Mar 14, 2009 at 1:54 PM, Sidar Ok <[email protected]> wrote: >> >>> @Fabio >>> I meant a look up, as in the pattern desc original, instead of an IoC. >>> Something that gives a singleton logger, nothing big. >>> >>> @Ayende >>> >>> Why do you think it will create a huge amount of work, care to elaborate >>> ? >>> >> >> NH does a lot of logging, all of which will have to be touched >> > > It is something, but not huge. > > NH logs _everything_. That means touching a lot of the code base. > >> >>> >>> Secondly, I don't think the abstraction is not required, first it will >>> make it possible to swap between loggers and log4net is not mandatory any >>> more, >> >> >> What is the point? I don't see any pain involve with this for our users. >> > > There are a lot of people using log4net just because NH (I am one of them) > > Again, what is the pain? Having an additional library? > >> >>> second, the configuration needed for Log4net, which is not a NH config, >>> will not be needed anymore, reducing unnecessary complexity and noise in >>> config. >>> >> >> Ha? You aren't going to change that, but instead of well known logging >> configuration style, you are not going to support N configuration styles. >> > > ?? Any configuration will not be mandatory, but logging config will be > outside NH's config. It is a good SoC, isn't it ? If user wants me to log, > (s)he can give me the logger (s)he can give me the one that's already used > within the app, nothing that I mandate. > Logging config is _already_ outside of NH. > >> I am sorry, but I am afraid that I don't see this providing any value for >> us. >> > > I respect your opinion, but if seeing no value doesn't mean objection I am > up for providing a patch for this. > Let me rephrase that, then. I see a very big change proposed, one that would be a breaking change, for something that doesn't really cause user any pain. In addition to that, I am not seeing any ROI of this change. That is my problem. Breaking changes should be avoided in general, and I don't really see a justification for doing something like this.
