log4perl-devel — Ongoing log4perl development

Re: [log4perl-devel] Re: logdie / logcroak

[<msc...@ao...>]

In a message dated Tue, 27 Aug 2002 2:47:57 AM Eastern Standard Time, er...@se... writes:

>Here ya go.

Hmm, this doesn't work. Could you do me a favor and run a 

    diff -Naur Log-Log4perl Log-Log4perl.erik

and resend the patch? This way I can apply it easily with "patch". Also, including the test case into the patch file would save me some work. Sorry for the inconvenience :)

>
>Also made a new patch file, as the tests caught a bug (woot!).

Nice, that's what test suites are for :)

>I also noticed that there's no turning off FATAL logging.

That's probably not intended, I think "OFF" should be higher prio than FATAL and therefore suppress it. Kevin, what do you think?

 -- Mike

Mike Schilli
log...@pe...

[log4perl-devel] Re: logdie / logcroak

[Erik S. <er...@se...>]

Here ya go.

Also made a new patch file, as the tests caught a bug (woot!). I forgot 
to wrap the various
incarnations of warn() around a $self->is_warn(), which means it would 
print out warnings
to STDERR but not log 'em.

I also noticed that there's no turning off FATAL logging. Not sure if 
that's a good thing or
not myself... although it certainly makes things like logdie() always 
work, regardless if you
turn off all logging whatsoever.

-e


Msc...@ao... wrote:

> In a message dated 8/16/02 11:36:06 PM Pacific Daylight Time, 
> er...@se... writes:
>
>
>> For consideration: a patch that has sundry logdie / logcroak and
>> similar. 
>
>
>
> Sorry for the long delay, Erik, just found some time to look at your 
> patch. Very nice! It's going to be a bit slower than "log && warn" 
> (another level on the call stack), but that's the price for 
> convenience. I'll apply the patch -- can you add a few test cases?
>
> -- Mike
> log...@pe... 

[log4perl-devel] Re: logdie / logcroak

[Erik S. <er...@se...>]

Sure thing.

-e

Msc...@ao... wrote:

> In a message dated 8/16/02 11:36:06 PM Pacific Daylight Time, 
> er...@se... writes:
>
>
>> For consideration: a patch that has sundry logdie / logcroak and
>> similar. 
>
>
>
> Sorry for the long delay, Erik, just found some time to look at your 
> patch. Very nice! It's going to be a bit slower than "log && warn" 
> (another level on the call stack), but that's the price for 
> convenience. I'll apply the patch -- can you add a few test cases?
>
> -- Mike
> log...@pe... 

[log4perl-devel] Re: logdie / logcroak

[<Msc...@ao...>]

In a message dated 8/16/02 11:36:06 PM Pacific Daylight Time, 
er...@se... writes:


> For consideration: a patch that has sundry logdie / logcroak and
> 

Sorry for the long delay, Erik, just found some time to look at your patch. 
Very nice! It's going to be a bit slower than "log && warn" (another level on 
the call stack), but that's the price for convenience. I'll apply the patch 
-- can you add a few test cases?

-- Mike
log...@pe... 

[log4perl-devel] logdie / logcroak

[Erik W.S. <er...@se...>]

For consideration: a patch that has sundry logdie / logcroak and
similar. I basically re-implemented some of Log::Agent's routines and
added my own touches. At it's core, these implement things such as
$logger->logFOO() for FOO in [warn, die, cluck, carp, croak, confess].

Cheers,
-e

[log4perl-devel] Log::Log4perl 0.22 released

[Mike S. <m...@pe...>]

Hi all,

just released Log::Log4perl 0.22 to CPAN, mainly featuring appender thresholds, a standard log4j
feature. Have fun!

-- Mike
--------------------------------------------------------
Mike Schilli log...@pe... http://perlmeister.com
--------------------------------------------------------

Re: [log4perl-devel] Log::Log4perl 0.21ready?

[Kevin G. <ke...@go...>]

Mike, I'm just sitting down at my desk.  Got a lot of catching up to
do.  Will get to this as soon as I can, but give me a day or two.


Msc...@ao... wrote:
> 
> Hi Kevin,
> 
> I churned out quite a lot of code while you were gone -- could you
> please take a look (it's all documented in Changes) and also run your
> benchmark again to figure out if I introduced any avoidable slowness?
> If everything's ok so far, I'd like to release 0.21 to CPAN.
> 
> -- Mike
> 
> Mike Schilli
> log...@pe...
> http://perlmeister.com
> http://log4perl.sourceforge.net

-- 
Happy Trails . . .

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

Re: [log4perl-devel] Log::Log4perl->init() using hash reference

[<msc...@ao...>]

In a message dated Fri, 9 Aug 2002 7:22:02 PM Eastern Standard Time, "Chris R. Donnelly" <cdo...@di...> writes:

>I do have a couple of suggestions though:

They're great, I've integrated them both, added test cases and documentation, and included you to the contributors list (let me know if that's ok). 

It's checked into CVS and will be released with 0.22 (check for log4perl.sourceforge.net on how to obtain a CVS snapshot, if you're eager to try it out :).

Thanks a bunch!

-- Mike

[log4perl-devel] Log::Log4perl->init() using hash reference

[Chris R. D. <cdo...@di...>]

Hi!  We have been using log4j in the Java arena for some time, and are glad
someone is finally taking on the perl port... we had been using
Log::Dispatch::Config until now, but it didn't do categories.  We appreciate
the effort!

I do have a couple of suggestions though:

1.  Would it be possible to pass other objects to init() other than a
filename or a reference to a string literal?  This would allow for
initialization of the log at runtime.  We have a program that forks
processes from a perl-based daemon, and each of the processes will write to
its own log.  Passing it a hash reference (or even better, an AppConfig
object) would enable us to do that fairly easily.  I have included a
modified Config.pm that parses key/value pairs in a hash reference into the
@text array, but did not do an AppConfig one in case you had any issues with
relying on other Perl modules.

2.  We will be using a wrapper class to wrap log4perl so that if log4perl
(or log4j) changes, we only need to change the one class instead of
everywhere it is referenced (we do this with log4j in Java as well).  The
problem there is that the %F and %L parameters in PatternLayout no longer
report the correct file and line (they always report the wrapper class).

To solve that, I borrowed a solution from Log::Dispatch::Config: I made
$caller_level a package-level variable inside of Log::Log4perl::Appender
(instead of always passing 3 to the render() function), and I increment it
by one when my wrapper class is initialized.  This gets the proper values
for the filename and line number.  I have included the modified Appender.pm
as well (both are based off of version 0.21).

Feel free to modify the patches I made in case they don't flow right... let
me know if I can be of any help!

Thanks,
# Chris

______________________________________________________________________
Chris Donnelly                cdo...@dm...
Software Systems Architect
Digital Motorworks
Austin, TX                    (512) 692-1101

Re: [log4perl-devel] Yet another log4perl...

[<Msc...@ao...>]

In a message dated 8/8/02 1:18:29 PM Pacific Daylight Time, 
sl...@ge... writes:

> We noticed it (log4j) about a year or two ago, and ourselves
> embarked on a project to create a perlified version of it.

Wow, I wish we had made contact earlier! Actually, Kevin and I stumbled into 
each other when he announced his plans a while ago. Later I had a small code 
base and we started furiously extending it.

> So I'm wondering what is the state if log4perl.  I noticed
> everywhere you had the word "ALPHA"...is it fully working?

Depends on your definition of "fully" :). It comes with a pretty detailed 
tutorial and I'm currently putting the final strokes on an article (also 
tutorial-style) which I plan on publishing real soon. All the features 
described in there are working and stable. If there's issues with anything we 
know about, it's listed in the "Changes" file. 
All in all, I think we've got a pretty nice implementation, we've been just a 
bit conservative by labelling it "ALPHA" to make clear it's not widely 
deployed and tested yet. Both Kevin and I are currently integrating it into 
bigger projects, I think we're gonna re-label it as soon as we're confident 
that there's no more hickups.
The regression test suite has about 120 test cases now, covering a lot of 
detail, so I would consider it pretty stable.

> I am
> also wondering if there is any way we might be able to
> contribute?  We do have a working version that is based on
> the older Category API vs the newer "Logger" API (which
> 

We're definitely interested in your help. Is your implementation publically 
available from somewhere? I'd definitely be interested to take a look. Maybe 
you could do the same with ours (perldoc Log::Log4perl) and let us know what 
you think (use the version from CVS to get the latest scoop).

-- Mike

Mike Schilli
log...@pe...
http://perlmeister.com
http://log4perl.sourceforge.net

[log4perl-devel] Yet another log4perl...

[Joe S. <sl...@ge...>]

I'm really glad to see that someone ported over log4j, it is
definitely one of the best logging packages around.

We noticed it (log4j) about a year or two ago, and ourselves
embarked on a project to create a perlified version of it.
We've been using our version for about the year, and always
intended to release it to CPAN but the developer who was
working on it went off to New Zealand and we just didn't
have the time to put it up.  It was just a matter of time
before someone (like you) did the same and put it up on
CPAN.

So I'm wondering what is the state if log4perl.  I noticed
everywhere you had the word "ALPHA"...is it fully working?
Would it be safe for us to maybe switch over to it?  I am
also wondering if there is any way we might be able to
contribute?  We do have a working version that is based on
the older Category API vs the newer "Logger" API (which
I prefer).

Thanks,
Joe

--
Joe Slagel
Geospiza, Inc.
3939 Leary Way NW, Seattle WA 98107
(206) 633-4403
sl...@ge...

[log4perl-devel] Log::Log4perl 0.21ready?

[<Msc...@ao...>]

Hi Kevin,

I churned out quite a lot of code while you were gone -- could you please 
take a look (it's all documented in Changes) and also run your benchmark 
again to figure out if I introduced any avoidable slowness? If everything's 
ok so far, I'd like to release 0.21 to CPAN.

-- Mike

Mike Schilli
log...@pe...
http://perlmeister.com
http://log4perl.sourceforge.net

Re: [log4perl-devel] Re: log4perl-devel post from erik@selberg.com requires approval

[Erik W.S. <er...@se...>]

heh, totally understood. I'm the same way. :)

-e
-- 
				Erik Selberg
"I get by with a little help	er...@se...
 from my friends."		http://www.selberg.org

[log4perl-devel] Re: log4perl-devel post from erik@selberg.com requires approval

[<Msc...@ao...>]

In a message dated 8/4/02 10:59:10 PM Pacific Daylight Time, 
log...@li... writes:

> to approve or deny the request.
> 
Hey Erik,

just subscribed you manually (looks like you used a non-supported way of 
subscribing :).

Hey, btw, if you want to fix something in log4perl, please post a patch to 
the mailing list, we'll take care of merging it in. Opening the CVS to too 
many folks usually results in chaos, but once the volume of your 
contributions exceeds a certain level, we're certainly gonna give you admin 
rights. Not that we generally distrust you, we're just careful with people we 
don't know :)

-- Mike

Mike Schilli
m...@pe...
http://perlmeister.com

[log4perl-devel] Re: multiple newlines

[<msc...@ao...>]

> Another possible nit:my formats tend to have %n in them for
> auto-newlines, as sometimes Iput in the newline, 
> sometimes I don't. Would it make sense for thevarious 
> appenders and whatnot to chomp the log message so that 
> therearen't dual newlines?

We don't control the appenders directly, we're using what's available in the Log::Dispatch namespace. The question is, though, should PatternLayout.pm check for a trailing newline and, if it's already there, *not* append another one if the format contains a %n? That's probably nice-to-have. I'll throw it in.

-- Mike

[log4perl-devel] Re: $logger->log() and multiple params

[<msc...@ao...>]

In a message dated Sat, 3 Aug 2002 3:05:55 PM Eastern Standard Time, er...@se... writes:

>First off, great port! 

Thanks! :)

>I have one suggestion: 

Funny, I noticed this one today -- that's a bug, the logger functions should take any number of arguments. I'll apply a fix today, thanks for the note!

-- Mike

[log4perl-devel] Re: "Best practices" for definining log4j categories?

[<Msc...@ao...>]

In a message dated 8/2/02 5:57:57 AM Pacific Daylight Time, 
Yoa...@mp... writes:

>  We've always stuck with the very basic 
> Logger logger = Logger.getLogger(MyClass.class.getName())
> 

That's what I thought -- it's probably too confusing to introduce a 2nd 
hierarchy unrelated to the class design. That's very valuable information for 
me, it's great that you've got such extensive experience using log4j in large 
projects. Thanks!

-- Mike

Mike Schilli
log...@li...

[log4perl-devel] "Best practices" for definining log4j categories?

[<Msc...@ao...>]

Hi there,

we in the log4perl team (log4perl.sourceforge.net, a Perl port of log4j) are 
wondering if there's any "best practices" for defining log4j categories. 
Sure, from what I've read so far, categories are usually resembling the class 
hierarchy of the software the logging code is used by. 

But are there any other recommendations or experiences shared by people 
having used log4j in huge projects?

Would be great if you could provide pointers or share your own experience.

-- Mike

Mike Schilli
m...@pe...
http://log4perl.sourceforge.net

[log4perl-devel] multiple newlines

[Erik W.S. <er...@se...>]

Hey guys,

Another possible nit:

my formats tend to have %n in them for auto-newlines, as sometimes I
put in the newline, sometimes I don't. Would it make sense for the
various appenders and whatnot to chomp the log message so that there
aren't dual newlines?

-e


-- 
				Erik Selberg
"I get by with a little help	er...@se...
 from my friends."		http://www.selberg.org

[log4perl-devel] $logger->log() and multiple params

[Erik W.S. <er...@se...>]

Hey guys,

First off, great port! Just getting my feet wet with it and it's
looking much better than my crappy three-line _log function I ripped
somewhere.

I have one suggestion: my existing _log function looks like:

my $v; # global verbosity string
sub _log {
  my $l = shift;
  print @_ if ($v >= $l);
}

I noticed that log looks like:

sub log {
# external api
##################################################
    my ($self, $priority, $message) = @_;

    croak "priority $priority isn't numeric" if ($priority =~ /\D/);

    my $which = Log::Log4perl::Level::to_level($priority);

    $self->{$which}($self, $message, Log::Log4perl::Level::to_level($priority));

}

Thus, $logger->warn("foo", "bar", "baz") will only print "foo".

Might a more general way to handle things look like:

sub log {
# external api
##################################################
    my $self = shift;
    my $priority = shift;
    my $message = join("", @_);

    croak "priority $priority isn't numeric" if ($priority =~ /\D/);

    my $which = Log::Log4perl::Level::to_level($priority);

    $self->{$which}($self, $message, Log::Log4perl::Level::to_level($priority));

}

?

Thanks,
-e

-- 
				Erik Selberg
"I get by with a little help	er...@se...
 from my friends."		http://www.selberg.org

[log4perl-devel] question

[Kevin B. <ke...@mo...>]

Hey, can log4perl use a SocketAppender to send log output to a Java 
socket (log4j)?

K

[log4perl-devel] Re: Log-Log4perl-0.17 missing Makefile.PL

[<msc...@ao...>]

In a message dated Tue, 16 Jul 2002 2:46:01 PM Eastern Standard Time, Mike Castle <da...@ix...> writes:

>FYI:
>Just tried to build 0.17 and noticed it was missing Makefile.PL.

Hi Mike,

thanks for the note. Sorry, I goofed with the uploads today, can you please check out 0.19 from 
    http://log4perl.sourceforge.net
(it's not on CPAN yet) and let us know if it works for you?

Thanks,

-- Mike

Mike Schilli
m...@pe...

[log4perl-devel] Test Message to log4perl-devel@lists.sourceforge.net

[Mike S. <m...@pe...>]

Hi Kevin,

just created log...@li... for bug reports, patches and suggestions. Both you
and me are on it and have admin rights, I'll send you the passwd separately ....

-- Mike
--------------------------------------------------------
Mike Schilli m...@pe... http://perlmeister.com
--------------------------------------------------------