Re: CharSequence messages

Created a JIRA issue for this: https://issues.apache.org/jira/browse/LOG4J2-1418 On Fri, Jun 3, 2016 at 7:27 PM, Mikael Ståldal wrote: > That was an idea, not something already in the code. > > On Fri, Jun 3, 2016 at 6:53 PM, Gary Gregory >

Re: CharSequence messages

That was an idea, not something already in the code. On Fri, Jun 3, 2016 at 6:53 PM, Gary Gregory wrote: > On Fri, Jun 3, 2016 at 9:10 AM, Gary Gregory > wrote: > >> Talk about NOT object-oriented though! :-( >> > > It feels like more and more

Re: CharSequence messages

On Fri, Jun 3, 2016 at 9:10 AM, Gary Gregory wrote: > Talk about NOT object-oriented though! :-( > It feels like more and more code like this instanceof has been sneaking in for 2.6 :-( A code review might help at some point, not today for me and not a blocker for 2.6.1.

Re: CharSequence messages

If provide a way for custom Logger:s to get access to MessageFactory2, then we would not need such instanceof check in AbstractMessageFactory, since the custom logger could invoke MessageFactory2.newMessage(CharSequence) directly. On Fri, Jun 3, 2016 at 6:10 PM, Gary Gregory

Re: CharSequence messages

Talk about NOT object-oriented though! :-( On Jun 3, 2016 3:49 AM, "Mikael Ståldal" wrote: > Would it make sense if AbstractMessageFactory was changed so that this > method: > > @Override > public Message newMessage(final Object message) { > return new

Re: CharSequence messages

I am thinking about how a custom Logger implementation can leverage the new functionality in 2.6, specifically the Scala API in LOG4J-1181 branch. Logger.getMessageFactory() return MessageFactory, even though the sole implementation in AbstractLogger do have a MessageFactory2 available. It would

Re: CharSequence messages

I don't see any problem with that. Let's make a similar change then also in ReusableMessageFactory. Sent from my iPhone > On 2016/06/03, at 19:49, Mikael Ståldal wrote: > > Would it make sense if AbstractMessageFactory was changed so that this method: > @Override >

CharSequence messages

Would it make sense if AbstractMessageFactory was changed so that this method: @Override public Message newMessage(final Object message) { return new ObjectMessage(message); } did this: @Override public Message newMessage(final Object message) { if (message instanceof CharSequence) {