From pjf at perltraining.com.au  Tue Nov  3 22:37:46 2009
From: pjf at perltraining.com.au (Paul Fenwick)
Date: Wed, 04 Nov 2009 17:37:46 +1100
Subject: [Melbourne-pm] Next meeting: 11th Novermber - 5.12 bug triage!
Message-ID: <4AF1213A.4040507@perltraining.com.au>

G'day Melbourne.pm,

Our next meeting is just over a week away, on November 11th, and I'm going
to hijack it for my own nefarious purposes.

Actually, I'm going to hijack it so that we can get Perl 5.12 out the door
by Coptic Christmas[1].

Jesse Vincent, the new Perl 5.12 pumpking, has asked me to help manage bug
triage in preparation for the 5.12 release.  In short, we need to figure out
if there's any showstopping bugs that have to be fixed in 5.12.  Helping
sort other bugs, pushing patches to p5p, merging duplicates, and closing
bugs that have been fixed (or aren't bugs in the first place) are all
bonuses.  None of this is particularly hard, and most of it doesn't require
any special access or special knowledge, provided that you already know Perl.

The hard part is inspecting 1,500 bug reports.

I'm in the process of drafting a plan that will mean that there won't be a
need to walk through 1,500 bug reports again.  I'll let you all know how
that works when I've figured that out myself.  :)

I'm not expecting Melbourne.pm to knock over 1,500 bug triages.  I am,
however, expecting us to make a decent dent.  I'm also hoping to turn the
bug triage into something of a competition between Perl Mongers groups, and
quite frankly, I'd like us to win.  This gives us a head start[2].  ;)

More details to follow later.

Looking forward to seeing you all there,

	Paul

[1] 7th January 2010.  Actually getting 5.12 released by then will be a
push, but we need a release deadline to plan all the other ones.  Actual
results may vary.

[2] I'm also hoping that so many people will be outraged by my blatant
favouritism of Melbourne.pm that they'll organise their own bug-triage
events. ;)

-- 
Paul Fenwick <pjf at perltraining.com.au> | http://perltraining.com.au/
Director of Training                   | Ph:  +61 3 9354 6001
Perl Training Australia                | Fax: +61 3 9354 2681

From thogard at abnormal.com  Wed Nov  4 18:11:43 2009
From: thogard at abnormal.com (Tim Hogard)
Date: Thu, 5 Nov 2009 02:11:43 +0000 (UTC)
Subject: [Melbourne-pm]  Multilanguage programmer neeed
In-Reply-To: <4AD67BFD.1040201@perltraining.com.au>
Message-ID: <200911050211.nA52Bh6Y017551@v.abnormal.com>


 We are looking for a full time Programmer for exciting projects
 in Canterbury and/or Clayton near the campus of Monash University.

Minimum requirements:
* Understanding and experience with web technologies
* Web user interface design
* Understanding and experience with FreeBSD/Solaris/Linux
* Understanding of embedded system issues
* Understanding of both procedural and OO design
* Ability to write code to standards that can be auditable
* Borne Shell scripting, C, Perl

Advantageous to have:
* Java with GWT
* Javascript
* Apache Internals
* Django
* Database using Berkeley DB or PostgreSQL
* Python
* Sysadmin knowledge
* Windows operating system knowledge
* Revision control experience

Some projects will require a current police check.

Please forward resume via email to umd at umd.com.au <mailto:umd at umd.com.au>
or post to:
Administration Manager
Unique Micro Design Pty Ltd
PO Box 4297
Mulgrave Vic 3170



From thogard at abnormal.com  Sun Nov  8 01:01:09 2009
From: thogard at abnormal.com (Tim Hogard)
Date: Sun, 8 Nov 2009 09:01:09 +0000 (UTC)
Subject: [Melbourne-pm] Modern Forum software?
Message-ID: <200911080901.nA8919qD058406@v.abnormal.com>


What are people using for forum software today?
It looks like most of the perl based ones are either 1) fast but have a dated interface or
2) are slow but look good.

I figure slashcode and its friends fit in with 1 and everything2 fits in catergory 2.

Is there something clean, lightweight and modern?

Also is there anything that doesn't require a SQL database?  If the
average post is about 500 characters and the minimum size for a
database is about 20 meg of memory, I figure you need about 40,000
messages before the database is faster than flat files.

-tim


From pml540114 at gmail.com  Sun Nov  8 06:05:55 2009
From: pml540114 at gmail.com (Peter Lawrence)
Date: Mon, 9 Nov 2009 01:05:55 +1100
Subject: [Melbourne-pm] Modern Forum software?
In-Reply-To: <200911080901.nA8919qD058406@v.abnormal.com>
References: <200911080901.nA8919qD058406@v.abnormal.com>
Message-ID: <196be9430911080605s289bea84lb2187b013fea13c9@mail.gmail.com>

On 08/11/2009, Tim Hogard <thogard at abnormal.com> wrote:
.
.
.
> Also is there anything that doesn't require a SQL database?  If the
> average post is about 500 characters and the minimum size for a
> database is about 20 meg of memory, I figure you need about 40,000
> messages before the database is faster than flat files.

As I recall, there is a Perl module out there that puts an SQL wrapper
around flat files. Why not track that down and use it to start with,
so you can upgrade to a "proper" database more easily as, when and if
the time comes? Or, for intermediate functionality and size, there is
a similar wrapper for the xbase file format. P.M.Lawrence.

From simon at unisolve.com.au  Sun Nov  8 14:07:07 2009
From: simon at unisolve.com.au (Simon Taylor)
Date: Mon, 09 Nov 2009 09:07:07 +1100
Subject: [Melbourne-pm] Modern Forum software?
In-Reply-To: <200911080901.nA8919qD058406@v.abnormal.com>
References: <200911080901.nA8919qD058406@v.abnormal.com>
Message-ID: <4AF7410B.8080403@unisolve.com.au>

Hello Tim,

> Is there something clean, lightweight and modern?
>
> Also is there anything that doesn't require a SQL database?  If the
> average post is about 500 characters and the minimum size for a
> database is about 20 meg of memory, I figure you need about 40,000
> messages before the database is faster than flat files.
>   

Have you reviewed YABB? Ive been using it here for almost a year, and 
I'm happy enough with it:

    http://www.anzaas.org.au/cgi-bin/yabb2/YaBB.pl

Cheers,

Simon

From ts at meme.com.au  Mon Nov  9 03:38:03 2009
From: ts at meme.com.au (Tony Smith)
Date: Mon, 9 Nov 2009 22:38:03 +1100
Subject: [Melbourne-pm] Modern Forum software?
In-Reply-To: <200911080901.nA8919qD058406@v.abnormal.com>
References: <200911080901.nA8919qD058406@v.abnormal.com>
Message-ID: <BB3AC6A3-984E-4F76-8D4C-E1209C7EDD5E@meme.com.au>

I'd meant to send this to the list but didn't notice that I had only  
replied direct to Tim until later, after which I would have let it  
ride had it not been for an accidental discovery today while some  
local malfunction* was choking my net connection.

It turned out that the TransForum home page was the only web page  
quick enough to load completely during that period. Even a simple  
Google search stalled indefinitely.

*power cycling my ADSL modem and Airport Extreme got rid of the problem.

----- start original message ------

On 08/11/2009, at 8:01 PM, Tim Hogard wrote:

> What are people using for forum software today?

I'm still using my 1998-99 TransForum 1.0 code with only the most  
minor anti-spam and cosmetic enhancements, in a large part because  
I've been waiting for Perl 6 grammars.

> It looks like most of the perl based ones are either 1) fast but  
> have a dated interface or
> 2) are slow but look good.

If you consider deeply indented threads to be a dated interface or  
consider a single comment level which forces commenters to say who  
they are trying to address looks good, then I'd suggest you are asking  
for a religious argument

> I figure slashcode and its friends fit in with 1 and everything2  
> fits in catergory 2.

I'm in awe of what Slashdot has managed of quite recent times, but  
don't see everything2 falling within my definition of a forum.

> Is there something clean, lightweight and modern?
>
> Also is there anything that doesn't require a SQL database?  If the
> average post is about 500 characters and the minimum size for a
> database is about 20 meg of memory, I figure you need about 40,000
> messages before the database is faster than flat files.

TransForum was/is written to perform well with limited resources and  
totally within the file system, relying on always appending to avoid  
the overwrite problems that broke Matt's forum under heavy load. It  
has its own in-built templating system which means it can look like  
almost anything. (We later turned that part into modules which we  
actively used for client applications for several more years but never  
retrofitted to TransForum itself.)

Right now I have some hopes that Wave is moving in the directions I  
wanted to go but time will tell as to how it is used in practice.




Tony Smith
0405 499 718
TransForum Developer
http://www.transforum.net/

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.pm.org/pipermail/melbourne-pm/attachments/20091109/190f9d1e/attachment.html>

From pjf at perltraining.com.au  Mon Nov  9 23:05:47 2009
From: pjf at perltraining.com.au (Paul Fenwick)
Date: Tue, 10 Nov 2009 18:05:47 +1100
Subject: [Melbourne-pm] The Great 5.12 bug triage (WEDNESDAY, 6pm)
Message-ID: <4AF910CB.7050404@perltraining.com.au>

G'day Melb.pm,

Tomorrow is a our Perl 5.12 bug triage.

    6:30pm, Wednesday 11th November (TOMORROW)
    Remasys Pty Ltd
    Level 1
    172 Flinders St
    (just opposite Federation Square)
    Map: http://preview.tinyurl.com/6sa7l8

This is a reminder that I'm co-opting you all to help triage bugs in
preparation for the release of Perl 5.12.  To save time on the night,
there's a few things I want to cover with you now.

Preparation:

* Laptops.  Bring 'em.  It will be hard to triage bugs without them.

* Bitcard accounts.  If you're already a CPAN author, you have one of
  these.   If not, go to https://www.bitcard.org/ and get one.

* Perl.  Preferably the latest stable release you can get.  We can
  close a number of bugs on old releases of Perl if they're no longer
  issues on new releases.

Our first target will be bugs on the Overview list:

	http://rt.perl.org/rt3/NoAuth/perl5/Overview.html

These are (roughly) all the tickets that have hit the bug queue but have not
yet been looked at by a human.

Our criteria for 5.12 show-stoppers are ANY of the following:

   * Anything with no work-around.
   * Anything that crashes Perl.
   * Anything that gives a wrong answer.
   * Anything security relevant.
   * Anything that's "easy" to fix.

However we'll also be trying to sort out any other bugs that we can.  There
are some meta-tickets that we'll be using for that task, and I'll go through
their use on the night.

Looking forward to seeing you all there!

	Paul

-- 
Paul Fenwick <pjf at perltraining.com.au> | http://perltraining.com.au/
Director of Training                   | Ph:  +61 3 9354 6001
Perl Training Australia                | Fax: +61 3 9354 2681

From toby.corkindale at strategicdata.com.au  Mon Nov  9 23:09:56 2009
From: toby.corkindale at strategicdata.com.au (Toby Corkindale)
Date: Tue, 10 Nov 2009 18:09:56 +1100
Subject: [Melbourne-pm] The Great 5.12 bug triage (WEDNESDAY, 6pm)
In-Reply-To: <4AF910CB.7050404@perltraining.com.au>
References: <4AF910CB.7050404@perltraining.com.au>
Message-ID: <4AF911C4.60405@strategicdata.com.au>

Paul Fenwick wrote:
> G'day Melb.pm,
> 
> Tomorrow is a our Perl 5.12 bug triage.

Also, I had previously volunteered to give a very short talk about Moose.

Cheers,
Toby

From pjf at perltraining.com.au  Mon Nov  9 23:16:33 2009
From: pjf at perltraining.com.au (Paul Fenwick)
Date: Tue, 10 Nov 2009 18:16:33 +1100
Subject: [Melbourne-pm] The Great 5.12 bug triage and Moose! (WEDNESDAY,
 6pm)
In-Reply-To: <4AF911C4.60405@strategicdata.com.au>
References: <4AF910CB.7050404@perltraining.com.au>
	<4AF911C4.60405@strategicdata.com.au>
Message-ID: <4AF91351.7050002@perltraining.com.au>

Toby Corkindale wrote:

> Also, I had previously volunteered to give a very short talk about Moose.

Tomorrow is Toby's most awesome talk on Moose, the post-modern object system
for Perl 5!  Toby is an internationally acclaimed conference speaker and
holds a Nobel Prize in software development!

After Toby's talk we'll be celebrating by triaging some bugs to help the
release of Perl 5.12.

	Paul


-- 
Paul Fenwick <pjf at perltraining.com.au> | http://perltraining.com.au/
Director of Training                   | Ph:  +61 3 9354 6001
Perl Training Australia                | Fax: +61 3 9354 2681

From toby.corkindale at strategicdata.com.au  Tue Nov 10 00:18:14 2009
From: toby.corkindale at strategicdata.com.au (Toby Corkindale)
Date: Tue, 10 Nov 2009 19:18:14 +1100
Subject: [Melbourne-pm] Projector - DVI?
Message-ID: <4AF921C6.1040303@strategicdata.com.au>

Oh, does the projector at Remasys have a DVI plug?
My macbook has a DVI output. (Well, in theory it has a VGA output too, 
but I don't have the adaptor.)

-Toby

From ddick at iinet.net.au  Tue Nov 10 01:14:44 2009
From: ddick at iinet.net.au (David Dick)
Date: Tue, 10 Nov 2009 20:14:44 +1100
Subject: [Melbourne-pm] Projector - DVI?
In-Reply-To: <4AF921C6.1040303@strategicdata.com.au>
References: <4AF921C6.1040303@strategicdata.com.au>
Message-ID: <4AF92F04.10408@iinet.net.au>

On 10/11/09 19:18, Toby Corkindale wrote:
> Oh, does the projector at Remasys have a DVI plug?
> My macbook has a DVI output. (Well, in theory it has a VGA output too,
> but I don't have the adaptor.)

No idea. i'll look into it.

From ts at meme.com.au  Tue Nov 10 13:18:53 2009
From: ts at meme.com.au (Tony Smith)
Date: Wed, 11 Nov 2009 08:18:53 +1100
Subject: [Melbourne-pm] Projector - DVI?
In-Reply-To: <4AF921C6.1040303@strategicdata.com.au>
References: <4AF921C6.1040303@strategicdata.com.au>
Message-ID: <157C1F99-7F0D-4999-AE55-A1C51E6D298F@meme.com.au>

On 10/11/2009, at 7:18 PM, Toby Corkindale wrote:

> Oh, does the projector at Remasys have a DVI plug?
> My macbook has a DVI output. (Well, in theory it has a VGA output too, but I don't have the adaptor.)

I'll have the VGA adaptor for my two year old MacBook with me. We know it works with the Remasys projector. Not quite so sure how wide a range of MacBooks it works with. I know it isn't interchangeable with the one I had earlier for a Mac Mini.




Tony Smith
Complex Systems Analyst
Meme Media
Melbourne, Australia
http://www.meme.com.au/

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.pm.org/pipermail/melbourne-pm/attachments/20091111/163432a3/attachment.html>

From toby.corkindale at strategicdata.com.au  Tue Nov 10 14:22:28 2009
From: toby.corkindale at strategicdata.com.au (toby.corkindale at strategicdata.com.au)
Date: Wed, 11 Nov 2009 09:22:28 +1100 (EST)
Subject: [Melbourne-pm] Projector - DVI?
In-Reply-To: <1445908258.21871257891718397.JavaMail.root@dmz03>
Message-ID: <1582181578.21891257891748491.JavaMail.root@dmz03>


----- "Tony Smith" <ts at meme.com.au> wrote:
> On 10/11/2009, at 7:18 PM, Toby Corkindale wrote:
> Oh, does the projector at Remasys have a DVI plug?
> My macbook has a DVI output. (Well, in theory it has a VGA output too,
> but I don't have the adaptor.)
> 
> 
> I'll have the VGA adaptor for my two year old MacBook with me. We know
> it works with the Remasys projector. Not quite so sure how wide a
> range of MacBooks it works with. I know it isn't interchangeable with
> the one I had earlier for a Mac Mini.

Thanks Tony.
I have a three year old MacBook.
I know that its external video port is not compatible with the adaptors I had for my previous apple laptop.. nor are they compatible with the adaptors for the latest apple lappies either.
(Mine is a Mini-DVI. The latest ones are Mini-DisplayPort.)
http://en.wikipedia.org/wiki/Mini-DVI <-- that's what mine looks like..

Cheers,
Toby

From pjf at perltraining.com.au  Tue Nov 10 14:48:52 2009
From: pjf at perltraining.com.au (Paul Fenwick)
Date: Wed, 11 Nov 2009 09:48:52 +1100
Subject: [Melbourne-pm] Projector - DVI?
In-Reply-To: <4AF921C6.1040303@strategicdata.com.au>
References: <4AF921C6.1040303@strategicdata.com.au>
Message-ID: <4AF9EDD4.7040105@perltraining.com.au>

G'day Toby / MPM,

Toby Corkindale wrote:
> Oh, does the projector at Remasys have a DVI plug?
> My macbook has a DVI output. (Well, in theory it has a VGA output too,
> but I don't have the adaptor.)

Just got a phone-call from David saying that an appropriate mini-DVI -> VGA
adapter has been located.  Many thanks to everyone who offered one!

Looking forward to seeing you all tonight,

	Paul

-- 
Paul Fenwick <pjf at perltraining.com.au> | http://perltraining.com.au/
Director of Training                   | Ph:  +61 3 9354 6001
Perl Training Australia                | Fax: +61 3 9354 2681

From ts at meme.com.au  Tue Nov 10 18:56:13 2009
From: ts at meme.com.au (Tony Smith)
Date: Wed, 11 Nov 2009 13:56:13 +1100
Subject: [Melbourne-pm] Snow Leopard and moving modules from 5.8 to 5.10
Message-ID: <CA710F8B-FDAD-4454-B74A-57AA193360B6@meme.com.au>

Bugger 5.12.

If anybody has a good method for getting the 5.10 equivalents of all my 5.8 CPAN modules so I don't have to keep thinking about which version I need to be running, please talk to me tonight.

Yeah I know it won't all happen automatically, but it would be nice if enough of it did that I can get my head around a once off clean up of whatever issues arise.




Tony Smith
0405 499 718
TransForum Developer
http://www.transforum.net/

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.pm.org/pipermail/melbourne-pm/attachments/20091111/af4fbf77/attachment.html>

From toby.corkindale at strategicdata.com.au  Tue Nov 10 19:08:02 2009
From: toby.corkindale at strategicdata.com.au (Toby Corkindale)
Date: Wed, 11 Nov 2009 14:08:02 +1100
Subject: [Melbourne-pm] Snow Leopard and moving modules from 5.8 to 5.10
In-Reply-To: <CA710F8B-FDAD-4454-B74A-57AA193360B6@meme.com.au>
References: <CA710F8B-FDAD-4454-B74A-57AA193360B6@meme.com.au>
Message-ID: <4AFA2A92.10601@strategicdata.com.au>

Tony Smith wrote:
> Bugger 5.12.
> 
> If anybody has a good method for getting the 5.10 equivalents of all my 
> 5.8 CPAN modules so I don't have to keep thinking about which version I 
> need to be running, please talk to me tonight.
> 
> Yeah I know it won't all happen automatically, but it would be nice if 
> enough of it did that I can get my head around a once off clean up of 
> whatever issues arise.


Method one:

cpanp
o
# wait for a while..
i 1..(maximum number)
# It'd probably be worth setting "automatically follow prerequisites" to 
Always before starting though..

Method two:
Check out 'perllocal.pod' on your system (probably exists in more than 
one place if your system is like mine).
That contains a list of all the modules you've installed..
So you can parse that and then hit up cpanp with the list.
.. I'm pretty sure there's a CPAN tool to do it for you, but I'm afraid 
it escapes me right now.

From alec.clews at gmail.com  Wed Nov 11 01:35:23 2009
From: alec.clews at gmail.com (Alec Clews)
Date: Wed, 11 Nov 2009 20:35:23 +1100
Subject: [Melbourne-pm] Perl 5.12 Bug triage process
Message-ID: <3c2b63c00911110135w7416b285qbf3eca6af0c3d5d9@mail.gmail.com>

Further to pjf's plans for world domination

The triage process is documented at
http://perl.net.au/wiki/Melbourne_Perl_Mongers#Perl_5.12_Bug_Triage

Please use and fix/improve.

Thanks

-- 
Alec Clews
Personal <alec.clews at gmail.com>             Melbourne, Australia.
Jabber:  alecclews at jabber.org.au             PGPKey ID: 0x9BBBFC7C
blog:http://alecthegeek.wordpress.com/

From tjc at wintrmute.net  Wed Nov 11 05:54:54 2009
From: tjc at wintrmute.net (Toby Wintermute)
Date: Thu, 12 Nov 2009 00:54:54 +1100
Subject: [Melbourne-pm] So many kinds of wrong..
Message-ID: <a20233a0911110554j76890dcpd573d9edcc36e14@mail.gmail.com>

I just wanted to point out this ticket, after tonight's bug triaging:
http://rt.perl.org/rt3/Ticket/Display.html?id=45091
How many kinds of wrong is this?!

CORE::exit is redefined as a GOTO, and then $SIG{__DIE__} is set to
call exit()...

The mind boggles..

I'm impressed at some of the bizarre corner-cases people can come up with.

-Toby

-- 
Turning and turning in the widening gyre
The falcon cannot hear the falconer
Things fall apart; the center cannot hold
Mere anarchy is loosed upon the world

From pat at patspam.com  Wed Nov 11 14:16:33 2009
From: pat at patspam.com (Patrick Donelan)
Date: Wed, 11 Nov 2009 17:16:33 -0500
Subject: [Melbourne-pm] Perl 5.12 Bug triage process
In-Reply-To: <3c2b63c00911110135w7416b285qbf3eca6af0c3d5d9@mail.gmail.com>
References: <3c2b63c00911110135w7416b285qbf3eca6af0c3d5d9@mail.gmail.com>
Message-ID: <42321ee20911111416y2ef3afc9r30201a7d1aa30904@mail.gmail.com>

How did the Bug Triage go? Is someone planning a Perl Iron Man blog
post(-mortem)?

I was bummed to not be able to make it, 6:30pm corresponded to approx
midnight new york time, but if it was a success then maybe the local perl
mongers over here can be convinced to run one of their own..

Patrick
http://patspam.com

On Wed, Nov 11, 2009 at 4:35 AM, Alec Clews <alec.clews at gmail.com> wrote:

> Further to pjf's plans for world domination
>
> The triage process is documented at
> http://perl.net.au/wiki/Melbourne_Perl_Mongers#Perl_5.12_Bug_Triage
>
> Please use and fix/improve.
>
> Thanks
>
> --
> Alec Clews
> Personal <alec.clews at gmail.com>             Melbourne, Australia.
> Jabber:  alecclews at jabber.org.au             PGPKey ID: 0x9BBBFC7C
> blog:http://alecthegeek.wordpress.com/
> _______________________________________________
> Melbourne-pm mailing list
> Melbourne-pm at pm.org
> http://mail.pm.org/mailman/listinfo/melbourne-pm
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.pm.org/pipermail/melbourne-pm/attachments/20091111/6979a571/attachment.html>

From tjc at wintrmute.net  Wed Nov 11 17:55:41 2009
From: tjc at wintrmute.net (Toby Wintermute)
Date: Thu, 12 Nov 2009 12:55:41 +1100
Subject: [Melbourne-pm] Moose examples
Message-ID: <a20233a0911111755w3d574b3bu5032203602389e93@mail.gmail.com>

Oh, I forgot to mention, I put a little Moose example set of files
online in case you wanted to look at a simple demo.
http://dryft.net/moose/

Toby

-- 
Turning and turning in the widening gyre
The falcon cannot hear the falconer
Things fall apart; the center cannot hold
Mere anarchy is loosed upon the world

From pjf at perltraining.com.au  Thu Nov 19 20:54:09 2009
From: pjf at perltraining.com.au (Paul Fenwick)
Date: Fri, 20 Nov 2009 15:54:09 +1100
Subject: [Melbourne-pm] Melbourne.pm bug triage report
Message-ID: <4B0620F1.6050506@perltraining.com.au>

G'day Melb.pm and p5p,

Here's the long overdue report on the great Melbourne Perl Mongers bug triage.

We managed to properly triage an estimated 68 tickets at Melbourne Perl
Mongers (many thanks to Myf White for calculating this), along with about 14
which would have been completely triaged if I had all the meta-tickets
properly set up.  That's about 82 tickets total, which is pretty darn
awesome for a single night and not the world's best wireless access[1].  A
huge thank-you to all the Melbourne.pm volunteers!

If you want to run your own triage event, then here's a few recommendations:

* Work in pairs or small groups.

* Make sure meta-tickets are well-known beforehand.
  On http://rt.perl.org/rt3/ these currently are:

	* RT 69710 - Perl 5.12.0 - Showstoppers
	* RT 70369 - Perl 5.12.0 - Expert assessment required
	* RT 70421 - Perl 5.12.0 - Non-critical
	* RT 70371 - Perl 5.005  - End of life
	* RT 70373 - Perl 5.6.x  - End of life

* Have someone on-hand to give people privileges.  Not everyone will have
  them sorted out beforehand.

* It's nice to have 5.10.1 floating around. A lot of old (5.6.x)
  bugs can just be closed if it's fixed in the most recent
  release.

* Break list of bugs into chunks and hand to each group.  This avoids
  double-triage of bugs.

* Go through an example or two as a group on a projector first, if you
  can.

All the best,

	Paul

[1] Federation square wifi allows outgoing http, but nothing else.
    No easy hopping on IRC.

-- 
Paul Fenwick <pjf at perltraining.com.au> | http://perltraining.com.au/
Director of Training                   | Ph:  +61 3 9354 6001
Perl Training Australia                | Fax: +61 3 9354 2681

From alecclews at gmail.com  Thu Nov 19 21:10:16 2009
From: alecclews at gmail.com (Alec Clews)
Date: Fri, 20 Nov 2009 16:10:16 +1100
Subject: [Melbourne-pm] Melbourne.pm bug triage report
In-Reply-To: <4B0620F1.6050506@perltraining.com.au>
References: <4B0620F1.6050506@perltraining.com.au>
Message-ID: <3c2b63c00911192110i5a9207e8ld6bdd2a26c4ef3fc@mail.gmail.com>

See also http://perl.net.au/wiki/Perl_5.X_bug_triage_process

Please improve as required

@pjf -- you may care to pass this to p5p as wel

2009/11/20 Paul Fenwick <pjf at perltraining.com.au>:
> G'day Melb.pm and p5p,
>
> Here's the long overdue report on the great Melbourne Perl Mongers bug triage.
>
> We managed to properly triage an estimated 68 tickets at Melbourne Perl
> Mongers (many thanks to Myf White for calculating this), along with about 14
> which would have been completely triaged if I had all the meta-tickets
> properly set up. ?That's about 82 tickets total, which is pretty darn
> awesome for a single night and not the world's best wireless access[1]. ?A
> huge thank-you to all the Melbourne.pm volunteers!
>
> If you want to run your own triage event, then here's a few recommendations:
>
> * Work in pairs or small groups.
>
> * Make sure meta-tickets are well-known beforehand.
> ?On http://rt.perl.org/rt3/ these currently are:
>
> ? ? ? ?* RT 69710 - Perl 5.12.0 - Showstoppers
> ? ? ? ?* RT 70369 - Perl 5.12.0 - Expert assessment required
> ? ? ? ?* RT 70421 - Perl 5.12.0 - Non-critical
> ? ? ? ?* RT 70371 - Perl 5.005 ?- End of life
> ? ? ? ?* RT 70373 - Perl 5.6.x ?- End of life
>
> * Have someone on-hand to give people privileges. ?Not everyone will have
> ?them sorted out beforehand.
>
> * It's nice to have 5.10.1 floating around. A lot of old (5.6.x)
> ?bugs can just be closed if it's fixed in the most recent
> ?release.
>
> * Break list of bugs into chunks and hand to each group. ?This avoids
> ?double-triage of bugs.
>
> * Go through an example or two as a group on a projector first, if you
> ?can.
>
> All the best,
>
> ? ? ? ?Paul
>
> [1] Federation square wifi allows outgoing http, but nothing else.
> ? ?No easy hopping on IRC.
>
> --
> Paul Fenwick <pjf at perltraining.com.au> | http://perltraining.com.au/
> Director of Training ? ? ? ? ? ? ? ? ? | Ph: ?+61 3 9354 6001
> Perl Training Australia ? ? ? ? ? ? ? ?| Fax: +61 3 9354 2681
> _______________________________________________
> Melbourne-pm mailing list
> Melbourne-pm at pm.org
> http://mail.pm.org/mailman/listinfo/melbourne-pm
>



-- 
Alec Clews
Personal <alec.clews at gmail.com>             Melbourne, Australia.
Jabber:  alecclews at jabber.org.au             PGPKey ID: 0x9BBBFC7C
blog:http://alecthegeek.wordpress.com/

From adrian at ash-blue.org  Thu Nov 19 22:58:33 2009
From: adrian at ash-blue.org (Adrian Masters)
Date: Fri, 20 Nov 2009 17:58:33 +1100 (EST)
Subject: [Melbourne-pm] Melbourne.pm bug triage report
In-Reply-To: <3c2b63c00911192110i5a9207e8ld6bdd2a26c4ef3fc@mail.gmail.com>
References: <4B0620F1.6050506@perltraining.com.au>
	<3c2b63c00911192110i5a9207e8ld6bdd2a26c4ef3fc@mail.gmail.com>
Message-ID: <3182.119.12.114.92.1258700313.squirrel@webmail.bella.lunarpages.com>

All,

>> We managed to properly triage an estimated 68 tickets at Melbourne Perl
>> Mongers (many thanks to Myf White for calculating this), along with about 14
>> which would have been completely triaged if I had all the meta-tickets
>> properly set up. ? That's about 82 tickets total, which is pretty darn
>> awesome for a single night and not the world's best wireless access[1]. ? A
>> huge thank-you to all the Melbourne.pm volunteers!

Awesome effort! Much kudos to all involved.

Adrian.




From jarich at perltraining.com.au  Mon Nov 23 04:52:13 2009
From: jarich at perltraining.com.au (Jacinta Richardson)
Date: Mon, 23 Nov 2009 23:52:13 +1100
Subject: [Melbourne-pm] Next meeting: 9th December with Damian Conway
Message-ID: <4B0A857D.909@perltraining.com.au>

G'day all,

As those who were at the last meeting will know, Damian has volunteered to give 
a talk at the December meeting.  As Damian is such a popular speaker, we are 
seeking alternatives to our usual venue, something able to seat 30-40 people 
would be a great start!  Paul has promised to approach Melbourne University's 
computer science department, but if you have your own contacts, or are able to 
volunteer to arrange a venue, that would make our lives much easier as we are 
currently swamped.

The meeting will be at the usual time:	6:30pm, 9th December

and if we don't find a new venue then it'll be standing room only at:

     Remasys Pty Ltd
     Level 1
     172 Flinders St
     (just opposite Federation Square)
     Map: http://preview.tinyurl.com/6sa7l8


The Missing Link by Damian Conway
=================================

What do duality, annoyances in yellow, chaos theory, tables and metatables, 
patterns and data, cinematic reboots, watching trees grow, code that writes code 
that writes code that writes code, debugging debuggers, eating your own dog, 
"objects-in, objects-out", successful failures, anti-stacks, and active null 
statements have in common? Watch as Damian weaves them together into a new and 
improbably useful module that demonstrates the awesome power and beauty of Perl 
5.10.


Depending on the venue, there may be pizza or other refreshments after the talk.

	J


From tjc at wintrmute.net  Mon Nov 23 18:00:52 2009
From: tjc at wintrmute.net (Toby Wintermute)
Date: Tue, 24 Nov 2009 13:00:52 +1100
Subject: [Melbourne-pm] The grass on the other side..
Message-ID: <a20233a0911231800ia9c8926reb675649eaaf1ce9@mail.gmail.com>

So I've been playing around with Scala lately, which is a
java-compatible language that can be roughly described as "Java, but
not sucking (as much)".

Anyway, I thought I'd give their web framework a try. Lift (ne? Scala
with Sails) is the usual MVC-like deal we've come to expect.

With Perl, Catalyst is the leading web framework, and people do tend
to complain that it installs half of CPAN to run.
I was interested to see what approach Lift took.. would be an
all-in-one package, like Jifty, or would it leverage lots of sub
projects? How does the Scala/Java ecosystem handle dependencies?

Well, initially it all seemed very easy. Mac OSX comes with maven
pre-installed, which meant I didn't even need to download or install
*anything* to try out Lift.
This confused me, and I spent a while trying to find out what to
download, sure that I needed to *do* something.
Eventually just ran the command from the tutorial, and presto, it went
off and downloaded a bunch of stuff for me. Ah.

Still, that's easier to get up and running than Catalyst. (which is
"cpanp i Catalyst::Devel .... wait a long time .... catalyst.pl
MyApp")

I then went to run the app immediately, thinking, ah, that was quick.
But it was at *this* point that it decided to automatically check and
update every dependency.

Hah! It had to happen eventually.. what felt like an age passed as
half of scala-tools.org was downloaded, and eventually the app
launched.

The thing that I found interesting was this choice for Lift to
auto-update your dependencies. I don't know how often it does it, I
know it isn't every time you subsequently run it.

How would that work in the Perl ecosystem?
I mean, imagine if Catalyst updated every single CPAN module on your
system, once a day?

I think we'd laugh at the concept.. It would be plagued by
incompatible updates hurting each other, or by one duff update
screwing the rest, or by circular dependencies getting stuck.
(As someone who has genuinely attempted to use cpanplus to auto-update
*every* installed module, regularly, I know it just doesn't work. It's
not hard to script it, but it's hard for it to not break your system
constantly.)


But apparently this is acceptable behaviour elsewhere. I'm curious
though.. does it suffer from the instability issues, or does it Just
Work?
If the Perl behaviour was changed to always-auto-update, would there
be a short period of pain, while we fixed stuff, but then a golden age
of reliability? Or is there something inherent in Perl's community
and/or release structure, that causes the pain?


Just musing..

Toby

From daniel at rimspace.net  Mon Nov 23 19:53:56 2009
From: daniel at rimspace.net (Daniel Pittman)
Date: Tue, 24 Nov 2009 14:53:56 +1100
Subject: [Melbourne-pm] The grass on the other side..
In-Reply-To: <a20233a0911231800ia9c8926reb675649eaaf1ce9@mail.gmail.com> (Toby
	Wintermute's message of "Tue, 24 Nov 2009 13:00:52 +1100")
References: <a20233a0911231800ia9c8926reb675649eaaf1ce9@mail.gmail.com>
Message-ID: <87hbskwpbf.fsf@rimspace.net>

Toby Wintermute <tjc at wintrmute.net> writes:

[...]

> The thing that I found interesting was this choice for Lift to auto-update
> your dependencies. I don't know how often it does it, I know it isn't every
> time you subsequently run it.

I understood that to be more a Maven (apt for Java) feature than a language
feature, when dealing with this sort of thing.

> How would that work in the Perl ecosystem?

Again, my understanding was that most projects listed their dependencies as
"random-library-1.1", and only version 1.1 would be installed, or suffice.

[...]

> I think we'd laugh at the concept.. It would be plagued by incompatible
> updates hurting each other, or by one duff update screwing the rest, or by
> circular dependencies getting stuck.

*nod*  Usually, in Java, updating arbitrary JARs is prone to these issues,
because the habit of distributing packages with applications, and using
versioned dependencies, means that many third party libraries don't practice
ABI compatibility.

[...]

> But apparently this is acceptable behaviour elsewhere. I'm curious
> though.. does it suffer from the instability issues, or does it Just Work?

My comments are based on my practical experience, save the Maven-ish bits,
where I only know a tiny bit.

> If the Perl behaviour was changed to always-auto-update, would there be a
> short period of pain, while we fixed stuff, but then a golden age of
> reliability?

Almost certainly not.

> Or is there something inherent in Perl's community and/or release structure,
> that causes the pain?

Nope.  Conversely, the Java community is *not* better at this than we are:
part of the reason that the habit of shipping third party binary libraries
with your application came about was to address the issue of ABI
compatibility.

Ruby fell into the same trap, with their gems system, if you are looking for
another reference point in a more dynamic (and Perl-like) language than Java.

        Daniel

-- 
? Daniel Pittman            ? daniel at rimspace.net            ? +61 401 155 707
               ? made with 100 percent post-consumer electrons

From damian at conway.org  Mon Nov 23 21:11:02 2009
From: damian at conway.org (Damian Conway)
Date: Tue, 24 Nov 2009 16:11:02 +1100
Subject: [Melbourne-pm] Next meeting: 9th December with Damian Conway
In-Reply-To: <4B0A857D.909@perltraining.com.au>
References: <4B0A857D.909@perltraining.com.au>
Message-ID: <832f158a0911232111w1d40a4c7u2b40625c5ca53aa9@mail.gmail.com>

Jacinta broke the bad news:

> As those who were at the last meeting will know, Damian has
> volunteered to give a talk at the December meeting.

Those at the last meeting may also recall that I have a classic Apple
Power Mac G4 Cube (http://bit.ly/7wVr2M) that is in need of a new home.

The specs of that luscious hunk of polycarbonate are:

    Processor: PowerPC G4 450MHz
          RAM: 256MB
         VRAM: 16MB
         Disk: 20GB
        Media: CD-RW
        Ports: 2 x USB1,
               2 x FireWire 400,
               1 x VGA,
               1 x ADC,
               1 x 10/100 ethernet,
               1 x 56k modem

I have previously managed to install MacOS 10.4 on it (and used it as a
media server), but it's currently running as-new with MacOS 9.1 instead.
It comes complete with prehistoric versions of iMovie, iTunes, Mail,
Quicktime, Netscape Communicator, Appleworks, IE5, Outlook Express 5,
and Palm Desktop already installed, plus the original installation CD.

It has no monitor, keyboard, or mouse, but can take standard USB input
devices and drive VGA or ADC monitors.

As competition for this postmodern masterpiece is likely to be intense,
we need some way to choose the lucky recipient. It was suggested that we
have the candidates arm wrestle for it on the night, but this seems
extremely ungeekly to me.

The keyboard is mightier than the bicep, so our contest shall be written, i
nstead of grunted. If you're interested in becoming the proud
owner of this ancient and mysterious device then you are invited,
between now and 9am on Monday 7th of December, to send me an essay
of up to 1000 words explaining why you deserve to receive it and what you
plan to do with it. In order to ensure your entry survives my spam filter,
please include the word "cube" somewhere in your subject line.

Points will be awarded for expression, style, originality, wit, pathos,
surprise, ruthless efficiency, a fanatical devotion to the Mac, and
anything else that amuses me. At the meeting on the 9th, we'll announce
the winner, read out their successful proposal, and present them with
their new G4 Cube. No correspondence will be entered into and the
judge's decision will be final. One entry per geek. Void where
prohibited by lore.

See you all next month,

Damian

From mathew.robertson at netratings.com.au  Tue Nov 24 16:15:07 2009
From: mathew.robertson at netratings.com.au (Mathew Robertson)
Date: Wed, 25 Nov 2009 11:15:07 +1100
Subject: [Melbourne-pm] The grass on the other side..
In-Reply-To: <87hbskwpbf.fsf@rimspace.net>
References: <a20233a0911231800ia9c8926reb675649eaaf1ce9@mail.gmail.com>
	<87hbskwpbf.fsf@rimspace.net>
Message-ID: <4B0C770B.3010608@netratings.com.au>


>> The thing that I found interesting was this choice for Lift to auto-update
>> your dependencies. I don't know how often it does it, I know it isn't every
>> time you subsequently run it.
>>     
>
> I understood that to be more a Maven (apt for Java) feature than a language
> feature, when dealing with this sort of thing.
>   
maven is not quite an "apt for java"... nobody *really* knows what it 
is, as it does weird stuff all over the place.  Apparently the 
developers think it is a project management tool, but is mostly used as 
a replacement for "ant" (aka "make" for java).

Its description file use lots of magic to determine when it does 
stuff... I've given up on numerous occasions trying to understand it -> 
I found it far too complicated to get anything useful done with it.
>> How would that work in the Perl ecosystem?
>>     
>
> Again, my understanding was that most projects listed their dependencies as
> "random-library-1.1", and only version 1.1 would be installed, or suffice.
>
> [...]
>   
When running a maven project in developer mode, maven will automatically 
download project-dependent packages+versions, and stick them in ~/.maven.

When you create the projects' release-build, normally the build system 
will package every dependency into a single jar/war/aar/mar/... file, so 
that production code doesn't auto-update.
>> I think we'd laugh at the concept.. It would be plagued by incompatible
>> updates hurting each other, or by one duff update screwing the rest, or by
>> circular dependencies getting stuck.
>>     
>
> *nod*  Usually, in Java, updating arbitrary JARs is prone to these issues,
> because the habit of distributing packages with applications, and using
> versioned dependencies, means that many third party libraries don't practice
> ABI compatibility.
>
> [...]
>
>   
>> But apparently this is acceptable behaviour elsewhere. I'm curious
>> though.. does it suffer from the instability issues, or does it Just Work?
>>     
>
> My comments are based on my practical experience, save the Maven-ish bits,
> where I only know a tiny bit.
>   
likewise...
>   
>> If the Perl behaviour was changed to always-auto-update, would there be a
>> short period of pain, while we fixed stuff, but then a golden age of
>> reliability?
>>     
>
> Almost certainly not.
>   
As with all code, at some point you will fix bugs or add features which 
are incompatible -> its just far too hard to maintain API and/or ABI 
compatibility.

In my experience, Perl's "version" handling is pretty good compared to 
other programming environments, but I'm in agreement -> there is no way 
(currently **) that the Perl community could auto-update version 
dependencies,  a given package may depend on another packages' 
version-specific API/ABI.

** If Perl's module support could be extended so that multiple module 
versions could be installed, when combined with Perl's version handling, 
it may be possible to do it... however, the Perl interpreter would also 
need to support having two versions of the same module loaded (aka each 
modules' symbols would also need version metadata - is this already 
available?)

>> Or is there something inherent in Perl's community and/or release structure,
>> that causes the pain?
>>     
>
> Nope.  Conversely, the Java community is *not* better at this than we are:
> part of the reason that the habit of shipping third party binary libraries
> with your application came about was to address the issue of ABI
> compatibility.
>   
Indeed - the Java community gave up trying to make a projects' 
dependencies play nice -> each project will now almost always ship the 
dependencies as binary blobs within the tarball.

Mathew Robertson
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.pm.org/pipermail/melbourne-pm/attachments/20091125/3f366eda/attachment.html>

From wayland at wayland.id.au  Wed Nov 25 12:33:45 2009
From: wayland at wayland.id.au (Timothy S. Nelson)
Date: Thu, 26 Nov 2009 07:33:45 +1100 (EST)
Subject: [Melbourne-pm] The grass on the other side..
In-Reply-To: <4B0C770B.3010608@netratings.com.au>
References: <a20233a0911231800ia9c8926reb675649eaaf1ce9@mail.gmail.com>
	<87hbskwpbf.fsf@rimspace.net> <4B0C770B.3010608@netratings.com.au>
Message-ID: <alpine.LRH.2.00.0911260732350.7574@gwalchavad.nelson.org.au>

On Wed, 25 Nov 2009, Mathew Robertson wrote:

> ** If Perl's module support could be extended so that multiple module versions could be
> installed, when combined with Perl's version handling, it may be possible to do it... however,
> the Perl interpreter would also need to support having two versions of the same module loaded
> (aka each modules' symbols would also need version metadata - is this already available?)

 	I think Perl 6 does at least the first, and possibly even the second.

---------------------------------------------------------------------
| Name: Tim Nelson                 | Because the Creator is,        |
| E-mail: wayland at wayland.id.au    | I am                           |
---------------------------------------------------------------------

----BEGIN GEEK CODE BLOCK----
Version 3.12
GCS d+++ s+: a- C++$ U+++$ P+++$ L+++ E- W+ N+ w--- V- 
PE(+) Y+>++ PGP->+++ R(+) !tv b++ DI++++ D G+ e++>++++ h! y-
-----END GEEK CODE BLOCK-----


From netwiz at crc.id.au  Mon Nov 30 02:31:19 2009
From: netwiz at crc.id.au (Steven Haigh)
Date: Mon, 30 Nov 2009 21:31:19 +1100
Subject: [Melbourne-pm] Handling UDP packets via IO::Socket::INET in a sane
	manner.
Message-ID: <38631838ebe5078b3a9d170cf1e1529e@localhost>

Hi all,

I'm wondering if anyone has a good method of using IO::Socket::INET in a
UDP proxy type environment...

I've written some (cruddy) code at the moment that listens on a UDP port
for an incoming packet. It then re-writes some details in the packet, and
then I need to send it on from the same UDP port number that we are
listening on to a different host.

Direction of traffic from client -> proxy -> server or server -> proxy ->
client can be done by inspecting the data in the packet so it isn't a huge
concern to have IO::Socket::INET have a sane way to tell traffic streams
apart, but it might be interesting (and easier) if it did.

>From my (somewhat incomplete) code at the moment, it seems that the UDP
packets are handled more like a TCP stream - in that even though I try to
set peeraddr and peerport before writing a packet to the socket, the packet
still ends up back at the client (which promptly gets confused!)

>From some extensive googling, it seems just about everything I've found
focuses on TCP streams - and my perl-foo is just not smart enough to filter
out the good stuff from the trivial example code to figure out what I need.

Any help would be appreciated!

-- 
Steven Haigh
 
Email: netwiz at crc.id.au
Web: http://www.crc.id.au
Phone: (03) 9001 6090 - 0412 935 897

From sam at nipl.net  Mon Nov 30 03:01:51 2009
From: sam at nipl.net (Sam Watkins)
Date: Mon, 30 Nov 2009 22:01:51 +1100
Subject: [Melbourne-pm] Handling UDP packets via IO::Socket::INET in a
	sane manner.
In-Reply-To: <38631838ebe5078b3a9d170cf1e1529e@localhost>
References: <38631838ebe5078b3a9d170cf1e1529e@localhost>
Message-ID: <20091130110151.GA17077@nipl.net>

On Mon, Nov 30, 2009 at 09:31:19PM +1100, Steven Haigh wrote:
> I'm wondering if anyone has a good method of using IO::Socket::INET in a
> UDP proxy type environment...

hi Steven,

I've got some UDP test code in perl, you're welcome to look at it if it helps,
and I'm happy to help you fix your proxy if I can.  Are you coding on Linux or
what OS?  I want to write some UDP convenience functions for C (my C dialect
brace), but that's another story!

http://sam.nipl.net/code/hacks/  <-- search for "udp" there are 6 files there
including "udp_relay" which sounds promising!  it's been a while since I worked
on these, so I'm not 100% sure if they actually did work, but I think at least
some of them did.

http://sam.nipl.net/contact.html  <--  please add me to yahoo messenger / jabber / irc / whatever
(anyone else can add me too, I like to chat sometimes!)

Sam


From netwiz at crc.id.au  Mon Nov 30 03:29:09 2009
From: netwiz at crc.id.au (Steven Haigh)
Date: Mon, 30 Nov 2009 22:29:09 +1100
Subject: [Melbourne-pm] Handling UDP packets via IO::Socket::INET in a
 sane manner.
In-Reply-To: <20091130110151.GA17077@nipl.net>
References: <38631838ebe5078b3a9d170cf1e1529e@localhost>
	<20091130110151.GA17077@nipl.net>
Message-ID: <66e68ec6e4d0728e8a7f011ee3ba7e4b@localhost>

On Mon, 30 Nov 2009 22:01:51 +1100, Sam Watkins <sam at nipl.net> wrote:
> On Mon, Nov 30, 2009 at 09:31:19PM +1100, Steven Haigh wrote:
>> I'm wondering if anyone has a good method of using IO::Socket::INET in
a
>> UDP proxy type environment...
> 
> hi Steven,
> 
> I've got some UDP test code in perl, you're welcome to look at it if it
> helps, and I'm happy to help you fix your proxy if I can.  Are you
coding on
> Linux or what OS?  I want to write some UDP convenience functions for C
(my C
> dialect brace), but that's another story!

It's on Linux. CentOS to be precise.

> http://sam.nipl.net/code/hacks/  <-- search for "udp" there are 6 files
> there including "udp_relay" which sounds promising!  it's been a while
since I
> worked on these, so I'm not 100% sure if they actually did work, but I
think at
> least some of them did.

I had a look at this, but udp_relay is written in *gasp* python!

> http://sam.nipl.net/contact.html  <--  please add me to yahoo messenger
/
> jabber / irc / whatever (anyone else can add me too, I like to chat
sometimes!)

Done.

-- 
Steven Haigh
 
Email: netwiz at crc.id.au
Web: http://www.crc.id.au
Phone: (03) 9001 6090 - 0412 935 897

From thogard at abnormal.com  Mon Nov 30 03:51:04 2009
From: thogard at abnormal.com (Tim Hogard)
Date: Mon, 30 Nov 2009 11:51:04 +0000 (UTC)
Subject: [Melbourne-pm] Handling UDP packets via IO::Socket::INET in a
	sane
In-Reply-To: <38631838ebe5078b3a9d170cf1e1529e@localhost>
Message-ID: <200911301151.nAUBp4wP079499@v.abnormal.com>

> 
> Hi all,
> 
> I'm wondering if anyone has a good method of using IO::Socket::INET in a
> UDP proxy type environment...
I will mention that this is in a SIP proxy hijack environment that Steve is
playing with.

I'll also add that I might be able to provide a SIP indial number
for people who want to play these types of games....  I've got a
few spare numbers I can route to the net or the Melbourne Wireless
network for people who want to play with that.

And if you call Steve you get something like....  "Thanks for
calling...all of our family members are currently with other
telemarketers but your call is important and will be placed in a
queue"...

-tim

From daniel at rimspace.net  Mon Nov 30 04:01:44 2009
From: daniel at rimspace.net (Daniel Pittman)
Date: Mon, 30 Nov 2009 23:01:44 +1100
Subject: [Melbourne-pm] Handling UDP packets via IO::Socket::INET in a
	sane manner.
In-Reply-To: <38631838ebe5078b3a9d170cf1e1529e@localhost> (Steven Haigh's
	message of "Mon, 30 Nov 2009 21:31:19 +1100")
References: <38631838ebe5078b3a9d170cf1e1529e@localhost>
Message-ID: <874ooc9q6v.fsf@rimspace.net>

Steven Haigh <netwiz at crc.id.au> writes:

> I'm wondering if anyone has a good method of using IO::Socket::INET in a
> UDP proxy type environment...
>
> I've written some (cruddy) code at the moment that listens on a UDP port
> for an incoming packet. It then re-writes some details in the packet, and
> then I need to send it on from the same UDP port number that we are
> listening on to a different host.
>
> Direction of traffic from client -> proxy -> server or server -> proxy ->
> client can be done by inspecting the data in the packet so it isn't a huge
> concern to have IO::Socket::INET have a sane way to tell traffic streams
> apart, but it might be interesting (and easier) if it did.
>
> From my (somewhat incomplete) code at the moment, it seems that the UDP
> packets are handled more like a TCP stream - in that even though I try to
> set peeraddr and peerport before writing a packet to the socket, the packet
> still ends up back at the client (which promptly gets confused!)

You didn't note how you were doing the packet receive and send operations;
traditionally, I would have expect this to work with the recv and sendto calls
as normal:

use Socket;
use IO::Socket::INET;
my $s = IO::Socket::INET->new(LocalPort => 9000, Proto => 'udp') or die;
my $buf;

my $dst = sockaddr_in(10000, inat_aton('192.168.1.1'));

while (1) {
    $buf = undef;
    $s->recv($buf, 65535) or last;
    $s->send($buf, undef, $dst) or die "sendto: $!";
}

Obviously, season to taste; I would personally grab something like AnyEvent to
help wrap up the process of buffering and asynchronous processing, unless you
don't care about buffering in the proxy.[1]

Presumably you can also fiddle the destination and use write, but I confess
that I have not tried: I grew up on recv and send(to) for UDP, and use them
automatically when needed.

> From some extensive googling, it seems just about everything I've found
> focuses on TCP streams - and my perl-foo is just not smart enough to filter
> out the good stuff from the trivial example code to figure out what I need.

Mostly, using the packet oriented API makes it work as expected, I find.

        Daniel

Footnotes: 
[1]  Given this is UDP you may, or may not, depending on your application.

-- 
? Daniel Pittman            ? daniel at rimspace.net            ? +61 401 155 707
               ? made with 100 percent post-consumer electrons

From netwiz at crc.id.au  Mon Nov 30 18:00:18 2009
From: netwiz at crc.id.au (Steven Haigh)
Date: Tue, 1 Dec 2009 13:00:18 +1100
Subject: [Melbourne-pm] Handling UDP packets via IO::Socket::INET in a
	sane manner.
In-Reply-To: <874ooc9q6v.fsf@rimspace.net>
References: <38631838ebe5078b3a9d170cf1e1529e@localhost>
	<874ooc9q6v.fsf@rimspace.net>
Message-ID: <F6252303-353E-4ABB-8D74-860DCDDFFD3E@crc.id.au>

On 30/11/2009, at 11:01 PM, Daniel Pittman wrote:
> You didn't note how you were doing the packet receive and send operations;
> traditionally, I would have expect this to work with the recv and sendto calls
> as normal:
> 
> use Socket;
> use IO::Socket::INET;
> my $s = IO::Socket::INET->new(LocalPort => 9000, Proto => 'udp') or die;
> my $buf;
> 
> my $dst = sockaddr_in(10000, inat_aton('192.168.1.1'));
> 
> while (1) {
>    $buf = undef;
>    $s->recv($buf, 65535) or last;
>    $s->send($buf, undef, $dst) or die "sendto: $!";
> }

Thanks for this Daniel, I've rewritten a lot of the socket code to sometime that seems to do what I want on the UDP side now - however I'm having some issues with (what is probably the SIP packet data) getting a response from the SIP registrar. I'm not 100% sure why - as it seems that the packet out is identical apart from addresses - but it's probably something simple and stupid that I've been blind too! ;)

I've put a copy up at http://www.crc.id.au/files/sip-filter.pl in case anyone would like to have a browse and help me hack away to get something workable.

--
Steven Haigh

Email: netwiz at crc.id.au
Web: http://www.crc.id.au
Phone: (03) 9001 6090 - 0412 935 897