Nice dig RB..

This conversation reminds me of why I've been questioning the reasoning behind 
integrating a logUI into the Log4J architecture.  That's what I've been struggling 
with recently.

My UI is, for all intents and purposes, feature-complete.  The churn on the receiver 
side is quite fast.  I'm worried if I implement something it won't mesh with what 
someone was thinking the day before, so I'll have to do rework - matching up with 
their vision.

The source side has been thought out and works well - except SocketAppender not 
supporting a layout..more proof that the reception side of things is not as far along 
as the source side.  I see a benefit to the plugin (receiver) configuration framework 
but that's it.

LoggingEvent has too much baggage - ThrowableInfo, LocationInfo - why aren't these 
just properties of the event?  That would fit better with other languages that don't 
support throwables or provide an equivalent to LocationInfo.  These definitely 
shouldn't be in the DTD!

Once I've reconciled myself with the existing integration issues I'll release my UI - 
maybe it's too flexible for some, maybe not enough for others, but it supports every 
feature I could think of for getting information from the log quickly and easily.

As a practical matter, I would much rather avoid the issue of synchronization and use 
some form of an immutable LoggingEvent object in my UI.  To support the tab panes and 
filering my UI supports requires a lot of examination of the events - and worrying 
about the partially constructed object problem isn't acceptable.  The synchronization 
required to prevent deadlock would definitely slow down the UI.  I think I read about 
immutability and the partially constructed object in a CS book somewhere.

So, accessors only in a LoggingEvent interface or a new ImmutableLoggingEvent that 
drops LoggingEvents off on it's way in my UI - I don't care.

There's my 2 cents.  If someone wants to provide constructive information on the 
direction of the receiver framework and cross-language support, please do.

Scott

-----Original Message-----
From:   [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
Sent:   Thu 3/20/2003 9:22 AM
To:     Log4J Developers List
Cc:     
Subject:        RE: LoggingEvent and XML transport protocol

"A level of indirection solves every problem in computer science"



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to