Re: POE HTTP server with WebSocket channel
Hey Guido,
I am the author of those modules. And yeah, they are broken and
experimental. I was attempting to create an unholy amalgamation of
MooseX::Declare and POE. It was an abomination that needed to be killed
with fire. I should make sure I clean up those modules.
That said, I had started hacking on a replacement stack on top of
Reflex, but lost interest. You're welcome to pick it up and finish it
if that is your cup of tea.
http://nickandperla.net/gitweb/?p=reflexive-role-psgiserver.git
It is based on top of Reflexive::Role::TCPServer
https://metacpan.org/pod/Reflexive::Role::TCPServer
If you have any questions, feel free to give me a shout on
irc.perl.org/#poe or #reflex
You can also contact me via XMPP at this same address.
On Thu, 12 Dec 2013 09:23:54 +0100
Guido Brugnara g...@leader.it wrote:
Il 12/12/2013 08:33, Konstantin Yakunin ha scritto:
Hello,
Try Plack::Builder
use Plack::Builder;
builder {
}
I like! Thanks.
I tried to install POEx-Role-PSGIServer from CPAN but points out the
lack of the package POEx-Role-SessionInstantiation [1]
Here ( https://rt.cpan.org/Public/Bug/Display.html?id=69745 ) the
author of [1] warns you deleted the package from CPAN because it is
no longer compatible with the latest versions of Moose :-(
So, The POEx-... packages are all considered experimental/broken?
What could I use as a PSGI server, along with POE in a single task?
gdo
2013/12/11 Guido Brugnara g...@leader.it mailto:g...@leader.it
===CUT===
The target is to listen on the same port:
http://myapp:8000/. and for websocket ws://myapp:8000/ws
I tried to use Protocol::WebSocket with the TCP socket used by
PoCo::Server::HTTP but my work is stalled.
I saw on CPAN that would be available the following packages :
POEx::Role::PSGIServer - Encapsulates core PSGI server behaviors
===CUT===
--
Nicholas Perez
XMPP/Email: n...@nickandperla.net
https://metacpan.org/author/NPEREZ
http://github.com/nperez
Re: poe and amqp/rabbitmq as a worker manager
None of this sounds particularly difficult, just a SMOP. The interesting bits aren't getting the job server up and running. What will be interesting is scaling the server. Am I correct to assume that the workers now directly pull from the queue? And then you would be changing that to a single queue puller that will then redistribute to worker processes? You may end up killing your over all capacity with an artificial bottleneck. We ended up building a similar system (but using 0mq, which has been both good and bad, but mostly bad, heh) at work with a very small/simple broker written in C for speed distributing work to individual workers. But instead of having complicated management, we just opted for a round robin approach. And instead of querying the broker for stats, we setup graphite and statsd and the worker processes themselves spit out metrics in a non-blocking fashion (after sending off the reply) so we can monitor things like request processing time, various counts on request resolution, etc. The workers themselves actually have a control channel to which we can individually address to tell them to pause/resume/stop. This gives us the flexibility to do rolling restarts of our workers when we are doing a deployment of updated worker code. And knowing that we can easily consume all of the cores on the machine, our infrastructure is such that we have multiple of such machines and do round-robin dns (with the ability to drop a host out for maintenance if needed). In the end, I would shy away from monolithic systems, especially in POE. POE is fast, especially with the right event loop, but ultimately, it is still a single process. Any time spent processing your management requests is time that it is not spending assigning work or gathering results from a worker. Some things to consider (and watch out for) if you decide to move forward with this: Inter-session communication is slow. Really slow. And the number of sessions that the kernel must keep track of also has an impact on performance (POE does a lot of book keeping). You'll need to figure out some sort of serialization mechanism for framing your requests to the workers. By default, POE::Filter::Reference uses Storable which is really slow and also produces pretty bloated output. You will need to possibly consider a different Filter module (I know of someone building a Filter using Sereal). Also be prepared to manage timeouts and your worker processes going away and what your master process will do in those cases. You will likely end up writing more error handling code than actual get-work-done code. For what it is worth, I actually wrote our worker code using Reflex (a better abstraction layer on top of POE). It basically spins up a singleton session and completes as many operations within a timeslice as possible to avoid going through the kernel as often. If latency in your worker processes is a concern you might consider using Reflex. There are various POE adaptors as well so they interoperate. On Thu, 14 Mar 2013 18:23:24 -0700 Kevin Goess kgo...@bepress.com wrote: We currently use rabbitmq for message between our web application and asynchronous workers. The worker management is somewhat ad hoc, and we're looking for a way to get a better handle on them. It looks like POE has the components that I want, so it's finally time for me to learn about POE. But there's an awful lot of POE material out there, and I'm afraid if I try to digest it all I'll have made a lot of false starts before I find the right path. Can anybody tell me if I'm going in the right direction, or if there's already something out there that does this? I think I want to use POE as a job server driven by the AMQP POE client, with workers in separate child processes handled by something like POE::Component::Daemon (which has a scoreboard) or POE::Wheel::Run. I'd like to be able to query the server on a management port with questions like - How many messages per queue are you receiving - What's the completion time for jobs on each queue - How idle/busy are your child workers? It should be able to take commands like add or drop these queues, and it should automatically take care of tasks like making sure no queue is being starved in favor of another queue. Is this the right idea? Is there a general direction for this that would be obvious to sketch out that would save me having to understand every example in the poe cookbook? Any pointers would be appreciated. Thanks! -- Nicholas Perez XMPP/Email: n...@nickandperla.net https://metacpan.org/author/NPEREZ http://github.com/nperez
Re: alarm function under while loop
On Fri, 2 Nov 2012 13:12:58 +0530 Gokul Prasad nhgokulpra...@gmail.com wrote: Hi, snip_giant_mess/ Can you nopaste a complete (non)working test case? Trying to help you debug this piece meal isn't working. Also, explain what it is you are trying to achieve? Perhaps you do not need to do something so complicated. -- Nicholas Perez XMPP/Email: n...@nickandperla.net https://metacpan.org/author/NPEREZ http://github.com/nperez
Re: POE, log4perl and sqlite
On Mon, 15 Oct 2012 10:11:32 +0300 Antti Linno antti.li...@gmail.com wrote: 1) Do POE's sessions run in parallel? Not unless you are forking. 2) Does POE::Component::Log4perl cure this problem? Unknown. We need to see some code. I could speculate on a number of things potentially causing this issue. Are you spinning up more than one process? Do you have any children processes that are running? Do you use any other mechanism to read from the sqlite file? What does your database code look like? Are you using alarm() (perl's alarm, not POE::Kernel's). Anyhow without any code it is difficult to debug. Please nopaste some code so we can help you. -- Nicholas Perez XMPP/Email: n...@nickandperla.net https://metacpan.org/author/NPEREZ http://github.com/nperez
Re: catching exceptions?
On Wed, 18 Apr 2012 15:57:27 +0300 Gabor Szabo ga...@szabgab.com wrote: How am I supposed to catch exceptions? Please see Exception Handling in POE::Kernel. -- Nicholas Perez XMPP/Email: n...@nickandperla.net https://metacpan.org/author/NPEREZ http://github.com/nperez
Re: How to debug POE based app and find memory leak
On Wed, 11 Apr 2012 11:33:53 +0300 Gabor Szabo ga...@szabgab.com wrote: Hi, I have written a small POE script fetching web pages but it is leaking a lot of memory. I wonder if you could give me advice how to track down this memory leak? In a more generic question, are there special tools to debug a program using POE ? regards Gabor It depends on where the leak might be. POE has fantastic facilities to tracing its own inner workings including sessions that might be sticking around. Take a look in POE::Kernel for turning on some of this information. From my own experience on debugging memory leaks in POE applications it is typically two places: sessions that don't get GC'd (zombies), and the heaps in those sessions with ever growing numbers of keys. After that, you might end up with a cyclic reference somewhere that doesn't get cleaned up, so you can use Devel::Cycle. Without looking the code, it is difficult to give you any more pointers. -- Nicholas Perez XMPP/Email: n...@nickandperla.net https://metacpan.org/author/NPEREZ http://github.com/nperez
Re: Multiple LWP::UserAgents at the same time
On Tue, 3 Apr 2012 16:57:59 +0300 Gabor Szabo ga...@szabgab.com wrote: my $ua = LWP::UserAgent::POE-new(timeout = 10); my $response = $ua-get($url) ; At first glance without reading the documentation to this module, I would say that this is wrong. Fetching the $url should be a non-blocking action that you give it a postback (or configured session/event to post) to notify you when a response is ready for reading. There is a proper HTTP client for POE that gives you this ability (albeit with a bit more verbosity). Take a look at POE::Component::Client::HTTP and see if it better fits the bill for what you need. -- Nicholas Perez XMPP/Email: n...@nickandperla.net https://metacpan.org/author/NPEREZ http://github.com/nperez
Re: Modules depending on different event loops in the same POE app?
It would probably make sense to run the RCU in a child and have it push events over the pipes of POE::Wheel::Run. Not as monolithic that way, but then you could generalize what kind of interface those two pieces use to chat. At that point you could then replace the RCU part with anything else that also understands the interface. On Sun, 21 Aug 2011 16:55:13 +0200 Gunnar Strand gurra.str...@gmail.com wrote: I'm working on a simple media control system and I am using POE with Tk. I found the RCU (remote control unit) perl module hierachy to use with my infrared remote control. RCU can be event-controlled, but it is using the Event event loop. Preferably I'd like to have everything controlled by the same event loop and not having to poll the RCU periodically. Is this doable? BR Gunnar -- Nicholas Perez XMPP/Email: n...@nickandperla.net http://search.cpan.org/~nperez/ http://github.com/nperez
Re: POE::Session major bug.
I do believe this is indeed a bug, but not in POE::Session. It seems
the GC semantics have changed somewhat and call() specifically invokes
a GC sweep in the cases where a return value is expected from an event.
I think this behavior is wrong (and I may have been the ultimate reason
this behavior exists at all [GC after call()], some long long time
ago sadly). For intra-session call()s, GC should not be called.
Effectively, the $kr_active_session is $session, meaning that we are
inside an event in $session that is current executing. GC should be
reserved for when that source event finishes, not in the middle of it.
So what I think is happening is this: his call invokes a GC sweep a
couple of times until it depletes the event_count at which point the GC
sweep will collect the session. But since kr_active_session is still
around, delay() has no problem enqueuing an event against it. Then when
the zombie session's state is invoked, Kernel happily dispatches even
when it can no longer resolve the session (for source_session), since
the reference is still valid. ASSERT_DEFAULT catches this, by the way.
Recommendation: remove the _data_ses_gc_sweep() calls in call().
For what it is worth the tests seem to run fine on my machine with
these calls removed from call().
On Fri, 8 Apr 2011 11:57:18 +0800
flw su2ad...@gmail.com wrote:
Sorry for my poor english, let me show it just by code.
flw@waker:~/study$ perl -MPOE -le 'print $POE::VERSION
$POE::Kernel::VERSION $POE::Session::VERSION'
1.299 1.299 1.299
flw@waker:~/study$ ./bug.pl
start
call bar
inner bar
call bar
inner bar
found bug.
call bar
call bar
call bar
flw@waker:~/study$ cat bug.pl
#!/usr/bin/perl
use strict;
use warnings;
use POE;
create POE::Session(
package_states = [
main = [qw( _start foo bar die_on_bug )],
],
);
run POE::Kernel();
sub _start{
print start\n;
$_[KERNEL]-yield( 'foo' );
}
exit 0;
my $found_bug;
sub foo {
my $sess = $_[KERNEL]-_resolve_session($_[SESSION]) || '';
if ( $_[SESSION] ne $sess and not $found_bug ){
$found_bug = 1;
print found bug.\n;
$_[KERNEL]-delay( die_on_bug = 3 );
}
print call bar \n;
my $ret = $_[KERNEL]-call( $_[SESSION] = bar = 1 );# bad
#$_[KERNEL]-call( $_[SESSION] = bar = 1 ) or 1; # bad
too #my @ret = $_[KERNEL]-call( $_[SESSION] = bar = 1 );# bad
too # $_[KERNEL]-call( $_[SESSION] = bar = 1 ); # good
$_[KERNEL]-delay( foo = 1 );
}
sub bar {
print inner bar\n;
}
sub die_on_bug {
exit 0;
}
flw@waker:~/study$
--
Nicholas Perez
XMPP/Email: n...@nickandperla.net
http://search.cpan.org/~nperez/
http://github.com/nperez
Re: POE and DBI database question
When the pre-packaged POE+DBI solutions don't work for you, you can always spin up a child that talks to you via STDIN/STDOUT using POE::Wheel::Run and performs the various DBI actions on your behalf. Then define a simple command/reply interface using dumb hashes and something like POE::Filter::Reference to transport between the child and parent process. And like Andy also said, you won't be able to deal with an ORM unless you do some fancy footwork to serialize the data into a digestible format and then reconstitute it across the process boundary. So you could use DBIx::Class in the child to perform the specific queries for you based on a set of parameterized Command objects you define. Then use HashRefInflator on the results to build a Result object and send it back down the wire. Something like this: Parent - Command - Filter - | - Filter - Result - Child | process boundary On Thu, 17 Feb 2011 21:07:40 +0100 Markus Jansen markus.jan...@ericsson.com wrote: Hi POE experts, is there someone who has used POE in an asynchronous way with DBI, without the necessity to reinvent the wheel, with some example or recipe at hand? To illustrate my problem: PoCo::SimpleDBI looks nice, but I am not too sure how well this behaves inside a forking environment (prefork server). DBIx::Connector claims to be great on the prefork matter, but I don't see an asynchronous POE example or recipe somewhere. DBIx::Class claims to be even greater for non-SQL addicts, and seems to have some code in common with DBIx::Connector, but again I fail how to integrate that best into a POE environment ... Any hints / comments are highly welcome! Best regards, Markus Markus Jansen Ericsson GmbH Eurolab RD Ericsson Allee 1 52134 Herzogenrath, Germany Phone +49 2407 575 5157 Fax +49 22407 575 150 Mobile +49 172 2742003 markus.jan...@ericsson.com www.ericsson.com Ericsson GmbH. Sitz: Düsseldorf. Registergericht: Amtsgericht Düsseldorf, HRB 33012. Geschäftsführer: Stefan Koetz (Vors.), Nils de Baar, Bernd Schmidt. Aufsichtsratsvorsitzender: Anders Runevad. This Communication is confidential. We only send and receive email on the basis of the terms set out at www.ericsson.com/email_disclaimerhttp://www.ericsson.com/email_disclaimer. -- Nicholas Perez XMPP/Email: n...@nickandperla.net http://search.cpan.org/~nperez/ http://github.com/nperez
Re: Session lifetime issue (post on behalf of)
On Tue, 9 Mar 2010 09:59:30 -0800 Zero Hero zeroh...@qoobly.com wrote: I'm trying to solve a common problem which occurs when I'm using PoCos. The crux of the issue is that a session doing a post to a PoCo, needs to stay alive for work done by a secondary post by that PoCo. Consider: 1. Session A posts to Session B 2. In Session B event handler, it posts to Session C on behalf of Session A. 3. Session B exits event handler. 4. Session A dies, because it's event was handled by B, but it has no way of knowing Session C is doing something for it. Since Session B typically belongs to the other persons PoCo (OPPC), I can't really get inside of it and change it. So the solution to this problem must not alter the code for the OPPC. This seems to be a very common second order problem. Is there some standard technique for solving this? I've looked at few PoCos for solutions, and searched for continuation like things (e.g. POE::Session::Yield::CC), but these don't seem to solve the problem. Thanks, Zerohero Typically what you can do is in Session B, increment the reference count for Session A (access via $_[+SENDER]) until after Session C notifies Session B that it's work is complete. Then you can decrement again. The other thing you could do is set up Session A to be persistent in some fashion, such as polling a shared result queue, where Session C will ultimately dump the result. -- Nicholas Perez XMPP/Email: n...@nickandperla.net http://search.cpan.org/~nperez/ http://github.com/nperez
Re: Execute Jobs parallel using JobQueue
On Thu, 3 Dec 2009 22:37:37 +0800 Ryan Chan ryanchan...@gmail.com wrote: On Wed, Dec 2, 2009 at 2:56 PM, Nick Perez n...@nickandperla.net wrote: On Wed, 2 Dec 2009 12:42:34 +0800 Ryan Chan ryanchan...@gmail.com wrote: If I goes with the fork solution, any abstraction recommended? I don't want to scare you off, but I can also suggest POEx::WorkerPool. -- unfortunately, make failed in CPAN in ubuntu 8.10 Can I get the failure output? What version of Perl? Also, why and operating system so ancient? -- Nicholas Perez XMPP/Email: n...@nickandperla.net http://search.cpan.org/~nperez/ http://github.com/nperez
Re: Execute Jobs parallel using JobQueue
On Wed, 2 Dec 2009 12:42:34 +0800 Ryan Chan ryanchan...@gmail.com wrote: If I goes with the fork solution, any abstraction recommended? I don't want to scare you off, but I can also suggest POEx::WorkerPool. -- Nicholas Perez XMPP/Email: n...@nickandperla.net http://search.cpan.org/~nperez/ http://github.com/nperez
Re: run an interactive command using POE::Wheel?
On Sun, 27 Sep 2009 10:59:48 +0800 woosley. xu. redic...@gmail.com wrote: $_[KERNEL]-sig_child($pid, got_child_signal); This is your problem line. Remove '$pid,' from the double quotes. -- Nicholas Perez XMPP/Email: n...@nickandperla.net http://search.cpan.org/~nperez/ http://github.com/nperez
Re: IKC bug for client spawn?
On Tue, 5 May 2009 10:34:53 -0700 Josh803316 josh803...@gmail.com wrote: It appears that if you start an IKC server and then an ikc client connects to it from within the same POE::Kernel-run() then we get the following error: 23289: Remote kernel 'host.domain-4a0076da5af9' already exists 23289: Remote kernel 'host.domain-4a0076da5af9' already exists Single program flow: POE::Component::IKC::Server-spawn(); POE::Component::IKC::Client-spawn(); POE::Kernel-run(); This is where we get the run time error from the client about the remote kernel already existing. The error doesn't seem to affect anything else but I assume this is either a bug or I'm doing something wrong. Erm, you do realize that IKC is mostly for inter process communication right? Running both within the same kernel (process) and it throwing that warning is /exactly/ what it should be doing. -- Nicholas Perez XMPP/Email: n...@nickandperla.net http://search.cpan.org/~nperez/ http://github.com/nperez
Re: IKC bug for client spawn?
On Tue, 5 May 2009 15:54:13 -0700
Josh803316 josh803...@gmail.com wrote:
Yes, thanks to some members of the list I have been taught that this
is indeed correct and not a bug. Makes more sense now.
I was trying to spawn a worker child but wasn't spawning it correctly
therefore the process was indeed inside the same kernel. I'm now
using a wheel::run to spawn the child process and then have the ikc
client connect. I'm sure there is a more elegant way to do this but I
need the portability of the clients spawning locally as different
processes, or on remote machines that would connect back to the
server.
(Inside the worker template package)
sub __spawn {
my $class = shift;
POE::Session-create(
inline_states = {
_start = \create_ikc_client,
worker_stdout = \worker_stdout,
worker_stderr = \worker_stderr,
worker_error = \worker_error,
},
args = [ $class ]
);
}
sub create_ikc_client{
my ( $kernel, $heap, $class ) = @_[ KERNEL, HEAP, ARG0 ];
my $child = POE::Wheel::Run-new(
Program = sub{
my ( $kernel, $heap ) = @_[ KERNEL, HEAP ];
# Wipe the existing POE::Kernel clean.
POE::Kernel-stop();
print STDERR Worker forked\n;
my $addr = $class-IP || '127.0.0.1';
my $port = $class-PORT || 5667;
my $trace = $class-TRACE || 0;
my $debug = $class-DEBUG || 0;
POE::Component::IKC::Client-spawn(
ip = $addr,
port = $port,
name = Client$$,
on_connect = sub {
POE::Session-create(
package_states =
## extract all
subroutines that don't start with __
## and allocate
them as states:
[
(__PACKAGE__)
= [ do {
no
strict 'refs';
grep {
!/^__/ and defined $_ } keys %{ __PACKAGE__ . :: };
} ],
($class)
= [ $class-SUBSCRIBE_METHODS ]
],
## pass args into
_start:
args = [$class],
options =
{ trace = $trace, debug = $debug },
);
},
on_error = sub {
my ( $kernel ) = $_[ KERNEL];
print STDERR Dumper($kernel);
}
);
# Run the new sessions.
POE::Kernel-run();
},
ErrorEvent = worker_error,
StdoutEvent = worker_stdout,
StderrEvent = worker_stderr,
CloseEvent = worker_close,
) or die $0: can't POE::Wheel::Run-new;
$heap-{worker} = $child;
return $child;
}
On Tue, May 5, 2009 at 3:41 PM, Nick Perez n...@nickandperla.net
wrote:
On Tue, 5 May 2009 10:34:53 -0700
Josh803316 josh803...@gmail.com wrote:
It appears that if you start an IKC server and then an ikc client
connects to it from within the same POE::Kernel-run() then we get
the following error:
23289: Remote kernel 'host.domain-4a0076da5af9' already exists
23289: Remote kernel 'host.domain-4a0076da5af9' already exists
Single program flow:
POE::Component::IKC::Server-spawn();
POE::Component::IKC::Client-spawn();
POE::Kernel-run();
This is where we get the run time error from the client about the
remote kernel already existing. The error doesn't seem to affect
anything else but I assume this is either a bug or I'm doing
something wrong.
Erm, you do realize that IKC is mostly for inter process
communication right? Running both within the same kernel (process)
and it throwing that warning is /exactly/ what it should be doing.
--
Nicholas Perez
XMPP/Email: n...@nickandperla.net
http://search.cpan.org/~nperez/ http://search.cpan.org/%7Enperez/
http://github.com/nperez
Wouldn't it be simpler to write the server, write
POE + Moose
So I had some expectations out of EC's effort, but doesn't look like the path being taken is one I necessarily want to follow. But it did give me some ideas on integrating POE and Moose. So I spent some time looking through various POE::Session implementations: strengths, weaknesses, and raw ideas. I really like Philip Gwyn's ideas on rolling up the POE stuff (KERNEL, SESSION, CALLER, etc) into a per-event object (POE::Session::PlainCall), for example. I also took some time to look at Chris Prather's efforts, which are very notable in that he appears to be wrapping POE::Session wholesale. But it seems to me that perhaps if a POE::Session implementation were actually a Moose::Role, we could all get succinctness we want, plus the mutability we need to redefine events at will. Apply the Role to any object to get an instant session. That said, I've started writing a simple POE::Session implementation (POE::Session::Moose: http://github.com/nperez/psm/tree/master). -- Nicholas Perez XMPP/Email: n...@nickandperla.net http://search.cpan.org/~nperez/ http://github.com/nperez
Re: Application Configuration Across Sessions
On Fri, 24 Apr 2009 19:32:11 -0500 johnr jo...@wowway.com wrote: Hello, What is the POE way of accessing application configuration across sessions? I was working with Config::General (storing config info in a session's heap). However, that is not accessible by other sessions. What is the approach for dealing with this? Thanks, John There are a couple of ways to do that. You could do some dependency injection on the constructed sessions so that they get their config info from the parent session. You could by convention, have each session know that the config information is stored in the parent session and therefore call $parent-get_heap(). You could have a lexical hash that is available to all of the sessions. Etc. Etc. I think you will find a myriad of ways that other people are doing it. But I believe the dependency injection of configuration information is the most common. -- Nicholas Perez XMPP/Email: n...@nickandperla.net http://search.cpan.org/~nperez/ http://github.com/nperez
RE: Automating network testing with POE (web services IPC)
It really sounds like you need to segregate the webapp from the actual daemon. And with that, you do not need to run POE within mod_perl (not that you could anyhow). But you obviously need POE communication between the web server and the POE daemon. So I recommend using POE::Component::IKC::ClientLite (http://search.cpan.org/~gwyn/POE-Component-IKC-0.2003/IKC/ClientLite.pm). As for worker/controller interaction, IKC can serve those needs too, without having to involve a protocol layer beyond a TCP socket and the serializer that IKC uses.
Re: which http server to use
Don't get me wrong, POE and perl in general is really fast considering that with a 800MHz Duron (I have ancient hardware, I know) I was able to manage 45+ requests per second, including streaming, against a database backend (SQLite) with my own home grown http server that incorporated HTML::Template::JIT, but it is still perl (even if bits wrap really tight C code). That was a really really specific application. For CGI execution you are better off sticking to something that speaks proper CGI like apache unless you are wanting to have a stand alone perl application (to ship as a web interface for some other POE enabled code or whatever). And even then you are probably better off rewriting your CGI scripts to exist within a POE environment which completely kills the idea of the common gateway interface. All in all it is going to end up being a lot of work without much gain (if any, comparatively for performance) if you already have a working solution under apache or something. The POE HTTP server solutions are only the bare essentials to manage a TCP connection and filter inbound data to build HTTP::Requests. You will need to write supporting code to get your CGI solutions integrated which may beyond the scope of your project. -- Nick Mathieu Longtin wrote: Hi, I'm trying to move a bunch of CGI scripts to a POE server, mostly for performance. Looking at CPAN, I see three different POE component HTTP server: POE::Component::Server::HTTP POE::Component::Server::SimpleHTTP POE::Component::Server::HTTPServer All of which have been patched in the last six months. So, I guess my questions are: - What are the differences, beside the names? - Which should I use? Thanks -Mathieu ___ Do you Yahoo!? Express yourself with Y! Messenger! Free. Download now. http://messenger.yahoo.com
Re: POE::Component::Jabber shoud support MUC
You are more than welcome to write an MUC module that makes use of the tools available in PCJ, but PCJ itself is only a transport mechanism. We are only building a connection broker. What you want to do with the connection after it is established is beyond the scope of the project. joe jiang wrote: i think the module should support MUC and have some good sample with it. Some message sync related project need this feature and i had to use thread instead of simple impletation of single stread.
Re: POE XML TCP Design
You should definitely use POE::Filter::XML. This is a shameless plug for my own code, but it really solves the problem of implementing a push parser that gives you meaningful chunks of data (POE::Filter::XML::Nodes) from which you can process. If you have any questions please let me know. Right now I am working on a new release and hope to have it on CPAN soon (the docs and the actual API are a teeny-tiny bit out of sync and I need to explain some behaviors a bit more), but what is out there right now is very stable and is in use by all all sorts of projects. -- Nick Lamb Joseph wrote: I want to know what components or wheels I need to use for my project. The project consist of simple IO::Socket(maybe POE client in the future) client sending XML data to POE Server. The Server parses the XML data and transforms the data to a HTML template file. Thank you, = Joseph Lamb __ Do you Yahoo!? New and Improved Yahoo! Mail - Send 10MB messages! http://promotions.yahoo.com/new_mail
[Fwd: Re: [PATCH] sslpoe-sanity-patch]
Oops. Reply went to Rocco but not the list: here it is. -- Nick ---BeginMessage--- I guess I should have sent my changes upstream which are in fact the very same changes here. I use this in PoCo::Jabber::Client::XMPP and also in Server::XMPP but adapted it for my own uses. I am not offering to maintain it, but I will test it against my code and perhaps share notes on these changes and others with whomever does take up maintainership. -- Nick Rocco Caputo wrote: On Sat, Aug 21, 2004 at 07:07:51PM -0400, Dylan William Hardison wrote: Some changes to SslClient.pm in sslpoe: Remove reference to $socket from %Filenum_Object. Reason: It was keeping the tied filehandle from being DESTROYed. In the READ method: Change die (handshake failed); to $! = 104; return undef. Reason: This makes POE handle it like a normal error. Dying allowed someone to send garbage via netcat and crash the program, thus adding a DoS to any server that uses SSL. That's not good. However, errno 104 is Connection reset by peer, and this is a lie. Really the handshake failed... Perhaps some other value should be used for $!. Added a DESTROY method that calls CLOSE, as CLOSE is not called otherwise. The patch is attached. KTHXAPPLIED! These changes are pretty trivial, and further work needs to be done on SslClient.pm to make it more robust. Also perhaps it should properly be named SSLSocketWraper or somesuch? The name is most confusing. I wouldn't mind that, but I don't do enough SSL work (as in ANY) to have a good idea what to do here. If you'd like, I'm ok with someone adopting and developing the project a bit. Preferably someone who uses it. ---End Message---
Re: POE scalability vs. forking
From my own experience using ithreads versus POE with Wheel::Run and without Wheel::Run with similar scripts that slam a server for requests with a uniprocessor machine, there isn't much difference other than POE provides a framework and some pretty slick pre-built components from the community that make very trivial work of writing load testers. Not exactly what you are looking for, but I guess we need to know what kind of hardware the tester will be running on to really tell you which will be better. nick Corey Goldberg wrote: POE can fork, look at the POE::Wheel::Run module or POE::Component::Child. These modules can do forking and management of those children. Thanks for the reply. But I am looking for the exclusive or ;) Either using forks (outside of POE), or using POE and making use of the event loop (without forking) -Corey --- Corey Goldberg [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED] ---
Re: Trouble with POE::Filter::XML
Reply to me off list and I'll help you out :)
I am the author
Aaron Craig wrote:
Hi all.
I would desperately love to use POE. All of the introductory material
makes it sound like just the sort of thing I need.
A bit of background. I'm writing a simple IRC chatbot script as a favor
to a friend, and as a learning experience. The requirements are simple
- get a bot into a channel and have it react appropriately to private
messages. Also, we want to be able to run more than one bot at a time.
Originally, I did a multi-threading script which works fairly well.
However, I'm having trouble with keeping the bot up in the channel. I'm
currently using Net::IRC, which is no longer supported. It has trouble
connecting to a server, and drops connections after a few hours. POE
seems to have a thriving community, and my friend, after asking around
on mailing lists and in IRC, got a lot of suggestions for
POE::Component::IRC.
So: CPAN -- install POE, et al.
The limited documentation at CPAN (which is, according to poe.perl.org,
the best place to look for documentation) was enough for me to get a
simple script up and running. The system seems fairly intuitive, and
some copy and modifying of Randal Shwartz' example code helped to get me
started.
Now, I want to be able to read in an XML configuration file for my
bots. CPAN -- install POE::Filter::XML. Check out the POD. To the
author's credit, he does admit that the documentation is as clear as
mud. Does anyone have any working example code to get a document tree
from an XML file with this module? :) The example in the synopsis of
the module's POD does not work, and an hour or so of playing around with
it hasn't gotten me any closer. I've tried to figure out what Wheels
do, especially ReadWrite, mostly by poring through the source, but I
just don't have my brain around the system enough to be able to figure
it out.
Is there documentation out there that I'm missing? I find it hard to
believe that such a useful system could have so little documentation.
I'm sure I've missed a link on poe.perl.org. Could someone point me in
the right direction.
Here's what I've got so far:
==
use strict;
use warnings;
use POE qw(Component::IRC Filter::XML);
my $sConfigPath = shift || chatbot.xml;
POE::Session-new
(
main = [ qw(_start _stop input_event) ],
);
POE::Kernel-run();
exit;
sub _start
{
print starting...\n;
open(XML, $sConfigPath) || die(Can't open config file (looking for
$sConfigPath): $!);
my $fltXML = POE::Filter::XML-new();
my $whlXML = POE::Wheel::ReadWrite-new
(
Handle= \*XML, # this isn't in POE::Filter::XML POD,
but Wheel complains if it isn't here (which, after all makes sense,
otherwise, what are you parsing?)
Filter= $fltXML, # verbatim (almost) from
Filter::XML
InputEvent= 'input_event', # verbatim (almost) from
Filter::XML
);
close XML; # is this the place to do this?
}
sub _stop
{
print done\n;
}
sub input_event
{
print got an input event!; # assuming later we'll shift()
interesting things off of @_ or something
}
==
Output from this script looks like this:
==
C:\perlperl chatbot.pl
starting...
done
rc LOOP VECTOR LEAK: MODE_RD = 0001
!!! Leaked fileno: 3 (total refcnt=2)
!!! Read:
!!! refcnt = 1
!!! ev cnt = 0
!!! session = POE::Session=ARRAY(0x1ba1b88)
!!! handle = GLOB(0x1a6b724)
!!! session = POE::Session=ARRAY(0x1ba1b88)
!!! event = POE::Wheel::ReadWrite(2) - select read
!!! Write:
!!! refcnt = 1
!!! ev cnt = 0
!!! session = POE::Session=ARRAY(0x1ba1b88)
!!! handle = GLOB(0x1a6b724)
!!! session = POE::Session=ARRAY(0x1ba1b88)
!!! event = POE::Wheel::ReadWrite(2) - select write
!!! Exception:
!!! refcnt = 0
!!! ev cnt = 0
==
It blocks after printing done
The !!! stuff gets printed after ctrl-C
Observant readers will now understand I'm on a Windows maching, 2k to be
precise. Perl is 5.8.3. I won't include -V output here, but will send
it later, if that seems necessary.
Ok, I think I've used enough bandwidth for now.
Thanks for the help.
