Incomplete log messages

bjhess · February 7, 2007, 1:20pm

I’m in the process of switching from Windows XP to OS X, and since
I’ve moved to OS X my log messages have been incomplete. This timing
may be coincidence, but somehow I doubt it.

Anyway, I’ve tried running my app locally under both webrick and
mongrel with the same results. All I see in my development.log is
something like the following:

DEBUG Wed Feb 07 06:14:37 -0600 2007 (2132)

INFO Wed Feb 07 06:14:37 -0600 2007 (2132)

So basically I have zero shot at debugging my app. Anyone know a way
to resolve this?

bjhess · February 8, 2007, 1:02pm

Anyone have thoughts on this?

Thanks!

bjhess · February 20, 2007, 9:05pm

For completeness, I’ll cross post my answer, which I posted in the
Deploying Rails group:

I’ve solved my problem. Thanks to everyone who read these ramblings -
sorry for bugging you all. Right up front, I did not share all of the
right information because frankly I did not know I had it.

For instance, I did not share that I had redefined the Logger’s
format_message method to improve my logger output. My definition was
as follows:

class Logger
def format_message(severity, timestamp, msg, progname)
“#{severity} #{timestamp} (#{$$})\n #{msg}\n”
end
end

It took me quite a while to figure out what was wrong with the above.
I tried some other folks’ format_message definitions with no luck:

http://dev.rubyonrails.org/ticket/2245

Then I found that Ruby 1.8.3 had transposed the msg and progname
arguments:

http://devblog.famundo.com/articles/2006/09/22/improving-the-rails-logger

Making the format look like what is depicted at the Wiki:

http://wiki.rubyonrails.com/rails/pages/HowtoConfigureLogging

Certainly this problem occurred because I went to a new Ruby version
on my new MacBook and coincidentally Dreamhost upgraded Ruby
recently. So the new format_message method is simply:

class Logger
def format_message(severity, timestamp, progname, msg)
“#{severity} #{timestamp} (#{$$})\n #{msg}\n”
end
end

-Barry