RE: [log4perl-devel] File compression and compiling

[Roger Y. <rog...@ey...>]

Kevin,

	I've attached a copy of the CompressedFile appender module that I've
written based on your recommendation below.  Please let me know if I've
missed anything.

	As for the change to the Log::Log4perl::Config module, here is the
diff -u:
$ diff -u Log-Log4perl-Conf*
--- Log-Log4perl-Config-ry.pm   2004-12-03 11:35:27.258550400 -0500
+++ Log-Log4perl-Config.pm      2004-12-03 11:28:14.626456000 -0500
@@ -461,8 +461,6 @@
         }
     }

-    $layout_class = ($layout_class !~ /::/) ?
"Log::Log4perl::Layout::$layout_class" : $layout_class;
-
     eval "require $layout_class" or
         die "Require to $layout_class failed ($!)";

	This ensures that regardless if the module is available or not, we
have the full package name for the layout.

	Yes, the change I made to the util module essentially disabled the
function.

Roger Yager


-----Original Message-----
From: Kevin M. Goess [mailto:ke...@go...] 
Sent: Monday, December 13, 2004 1:49 AM
To: rog...@ey...
Cc: log...@li...
Subject: Re: [log4perl-devel] File compression and compiling

Roger, thanks for your submission, sorry for the late response, it's 
been a busy week for me, and I suspect Mike Schilli is out of town.  A 
small point, diff -u is somewhat more traditional and easy to read than 
diff -c.

 >       Also, I have included file compression (using IO::Zlib) into the
 > Log::Log4perl::Appender::file (see diff below).

An interesting idea.  It looks to me like this changes the semantics of 
Appender::File to the point where it would be better off in its own 
appender class, like Appender::CompressedFile.  It clashes with "pipe", 
which you handle, as well as with "autoflush", which you don't.  And I 
wonder about the reliability/performance if the IO::Zlib uses the 
external gzip option.

> .         Modified Log::Log4perl::Config to handle abbreviated layout 
> names outside of the call to Log::Log4perl::Util::module_available (see 
> diff below).

I see from your diff where you took it out, but where did you put it 
back in?

> .         Modified Log::Log4perl::Util::module_available to always 
> return true  (see diff below).

Which disables the function, correct?

Mike, shouldn't the module_available check %INC first, in which case it 
doesn't have to try to load the module at all?


-- 
Happy Trails . . .

Kevin M. Goess
(and Anne and Frank)
904 Carmel Ave.
Albany, CA  94706
(510) 525-5217