php-java-bridge-users

[Php-java-bridge-users] PHP/Java Bridge 2.0 snapshot

[Jost B. <jos...@ya...>]

A binary snapshot of version 2.0 is available:

 http://php-java-bridge.sf.net/JavaBridge.jar
 http://php-java-bridge.sf.net/php

The server can be started with:

  java -jar JavaBridge.jar

The PHP binary is version 5.0.3 with the bridge
compiled in.

The official 2.0 version will be available next month.

Please report bugs to the mailing list or open a
ticket.  

Regards,
Jost Boekemeier



	

	
		
___________________________________________________________ 
Gesendet von Yahoo! Mail - Jetzt mit 250MB Speicher kostenlos - Hier anmelden: http://mail.yahoo.de

Re: [Php-java-bridge-users] PHP/Java Bridge 2.0 snapshot

[Jost B. <jos...@ya...>]

 --- Jost Boekemeier <jos...@ya...> schrieb: 
> A binary snapshot of version 2.0 is available:
> 
>  http://php-java-bridge.sf.net/JavaBridge.jar

I forgot to mention that you must use JDK 1.5 to run
the bridge.  The bridge currently does *not* run with
jdk1.4 or earlier.


Jost


	
		
___________________________________________________________ 
Gesendet von Yahoo! Mail - Jetzt mit 250MB Speicher kostenlos - Hier anmelden: http://mail.yahoo.de

Re: [Php-java-bridge-users] PHP/Java Bridge 2.0 snapshot

[Marek L. <ma...@le...>]

Jost Boekemeier napisał(a):
> A binary snapshot of version 2.0 is available:
> 
>  http://php-java-bridge.sf.net/JavaBridge.jar
>  http://php-java-bridge.sf.net/php
> 
> The server can be started with:
> 
>   java -jar JavaBridge.jar
> 
> The PHP binary is version 5.0.3 with the bridge
> compiled in.
> 
> The official 2.0 version will be available next month.
> 
> Please report bugs to the mailing list or open a
> ticket.  
> 
> Regards,
> Jost Boekemeier
Hello Jost,
After start of PHP I saw something like this in JavaBridge console:
lut 18 07:07:35 JavaBridge DEBUG: Request from unknown client

But this is only a warning. But, I've made simple test:


require_once "Benchmark/Timer.php";
$timer = new Benchmark_Timer();
$timer->start();
for ($i = 0; $i < 100; $i++) {
     $string = new Java("java.lang.String", "sdsdsd");
}
$timer->stop();
$timer->display();

The results:
-------------------------------------------------------------
marker    time index            ex time               perct
-------------------------------------------------------------
Start     1108710914.89464700   -                       0.00%
-------------------------------------------------------------
Stop      1108710917.28214400   2.38749718666         100.00%
-------------------------------------------------------------
total     -                     2.38749718666         100.00%
-------------------------------------------------------------

PHP bug, an object destructor was called after module shutdown



As you see there is some problem with object destructor and also 
JavaBridge is quite slow - over 2 seconds in making 100 strings objects 
- is this normal behavior ?


Regards,
ML

 

Re: [Php-java-bridge-users] PHP/Java Bridge 2.0 snapshot

[Marek L. <ma...@le...>]

Marek Lewczuk napisał(a):
> Jost Boekemeier napisał(a):
> 
>> A binary snapshot of version 2.0 is available:
>>
>>  http://php-java-bridge.sf.net/JavaBridge.jar
>>  http://php-java-bridge.sf.net/php
>>
>> The server can be started with:
>>
>>   java -jar JavaBridge.jar
>>
>> The PHP binary is version 5.0.3 with the bridge
>> compiled in.
>>
>> The official 2.0 version will be available next month.
>>
>> Please report bugs to the mailing list or open a
>> ticket. 
>> Regards,
>> Jost Boekemeier
> 
> Hello Jost,
> After start of PHP I saw something like this in JavaBridge console:
> lut 18 07:07:35 JavaBridge DEBUG: Request from unknown client
> 
> But this is only a warning. But, I've made simple test:
> 
> 
> require_once "Benchmark/Timer.php";
> $timer = new Benchmark_Timer();
> $timer->start();
> for ($i = 0; $i < 100; $i++) {
>     $string = new Java("java.lang.String", "sdsdsd");
> }
> $timer->stop();
> $timer->display();
> 
> The results:
> -------------------------------------------------------------
> marker    time index            ex time               perct
> -------------------------------------------------------------
> Start     1108710914.89464700   -                       0.00%
> -------------------------------------------------------------
> Stop      1108710917.28214400   2.38749718666         100.00%
> -------------------------------------------------------------
> total     -                     2.38749718666         100.00%
> -------------------------------------------------------------
> 
> PHP bug, an object destructor was called after module shutdown
> 
> 
> 
> As you see there is some problem with object destructor and also 
> JavaBridge is quite slow - over 2 seconds in making 100 strings objects 
> - is this normal behavior ?

One correction: the performance is low because of that I've run 
JavaBridge in second console which prints all debug informations. When 
I've run with " > /dev/null &" then the results looks much better:

-------------------------------------------------------------
marker    time index            ex time               perct
-------------------------------------------------------------
Start     1108711271.00199500   -                       0.00%
-------------------------------------------------------------
Stop      1108711271.10505000   0.103055000305        100.00%
-------------------------------------------------------------
total     -                     0.103055000305        100.00%
-------------------------------------------------------------


PHP bug, an object destructor was called after module shutdown


 

Re: [Php-java-bridge-users] PHP/Java Bridge 2.0 snapshot

[Jost B. <jos...@ya...>]

 
> PHP bug, an object destructor was called after
> module shutdown

Yes (it's a PHP bug, not a bug in the bridge). There
are a number of tickets regarding this problem.  I
don't know if or when the PHP people are going to fix
it.


> JavaBridge is quite slow - over 2 seconds in making
> 100 strings objects 

Depends on the network latency.  If you talk to the
server via a slow network with ~2ms latency, this is
feasible. For local sockets (either TCP or Unix
domain) the new protocol makes the bridge at least two
times faster.  From what I have seen the bench now
executes twice as fast as before: ~11s vs. ~4 seconds
native java code. 


> - is this normal behavior ?

Certainly.  Every RPC protocol has this problem. 
That's why one normally creates "value objects" on the
client-side and then sends these to the server. 
Bandwith is not the problem, latency is.

One could invent techniques such as "reply cookies" to
minimize the number of roundtrips.  But this is not
worth it, imho.  If you use local sockets, the
communication is reasonably fast.  If you want to talk
to a remote server, use value classes to communicate
with it.


Jost


	
		
___________________________________________________________ 
Gesendet von Yahoo! Mail - Jetzt mit 250MB Speicher kostenlos - Hier anmelden: http://mail.yahoo.de