cppcms-users

[Cppcms-users] Mounting both synchronous and asynchronous applications

[CN <cn...@gr...>]

Hello,

Imagine my site has both synchronous (for user login) and asynchronous  
(for interactive activities) pages. Is it a correct and good practice to  
mount both synchronous and asynchronous applications like the following  
code?

//code begin
class login:public cppcms::application{ .... }
class interaction:public cppcms::application{ .... }

int main()
{
cppcms::service s(argc,argv);

s.applications_pool().mount(cppcms::applications_factory<login>());

booster::intrusive_ptr<interaction> i=new interaction(s);
s.applications_pool().mount(i);

s.run();
}
//code end

Please also correct me if my following thoughts are wrong:

- When a request arrives, cppcms::service.run() loops all mounted  
applications and pass that request to every mounted application. It is the  
responsibility of every application to filter the interested passed in  
request by applying dispatcher().assign().

- This program can fire multiple instances of "login" class. Every login  
instance runs as a separate thread. Thus this program can concurrently  
process multiple incoming synchronous (login) requests.

- This program process asynchronous requests only one at a time because  
only one instance of class "interaction" is created. As such, asynchronous  
applications are most likely the bottleneck of performance when the load  
is heavier.

Thank you in advance!
CN

Re: [Cppcms-users] Mounting both synchronous and asynchronous applications

[Artyom <art...@ya...>]

> From: CN <cn...@gr...>

> Hello,
> 
> Imagine my site has both synchronous (for user login) and  asynchronous  
> (for interactive activities) pages. Is it a correct and  good practice to  
> mount both synchronous and asynchronous applications  like the following  
> code?
> 
> //code begin
> class login:public  cppcms::application{ .... }
> class interaction:public cppcms::application{  .... }
> 
> int main()
> {
> cppcms::service  s(argc,argv);
> 
> s.applications_pool().mount(cppcms::applications_factory<login>());
> 
> booster::intrusive_ptr<interaction>  i=new  interaction(s);
> s.applications_pool().mount(i);
> 
> s.run();
> }
> //code  end
> 


Yes and not, you should provide a mount point
that would distinguish between them:

See: 

- 
http://art-blog.no-ip.info/cppcms_ref_v0_99/classcppcms_1_1applications__pool.html

- http://art-blog.no-ip.info/cppcms_ref_v0_99/classcppcms_1_1mount__point.html


> Please also correct me if my following thoughts are wrong:
> 
> -  When a request arrives, cppcms::service.run() loops all mounted  
>   applications and pass that request to every mounted application. It is  the  

>   responsibility of every application to filter the interested  passed in  
>   request by applying dispatcher().assign().
> 

Not it would pass the request to the first application
that matches the mount point pattern. And it would not
it to only one such application.

Note it is multi-layered search:

If there is no specific patch assigned to dispatcher
it would return error 404 (unless you had redefined
application::main and do not use dispatcher

> - This  program can fire multiple instances of "login" class. Every login  
>   instance runs as a separate thread. Thus this program can concurrently  
>   process multiple incoming synchronous (login) requests.
> 

Yes and No.

A pool of synchronous applications is created.
At each request once instance is taken from the
applications pool and passed to thread pool for execution.

Once it completes it is recycled for reuse in the pool.

So multiple instances run synchronously concurrently.

> - This  program process asynchronous requests only one at a time because  
> only  one instance of class "interaction" is created.
> 

No, asynchronous application is executed in main event loop,
when new request is ready it is passed to the application, unlike
the case of synchronous application it can detach the http::context
object and "keep" it for further response (to provide long polling).
and then get more requests.

So basically asynchronous application can handle multiple
request and this is needed mostly for server side events
and long polling.

See examples/chat

> As such, asynchronous  
> applications are most likely the bottleneck of performance when the  load  
> is heavier.

If you do heavy processing there then yes, but you should not.
You are expected to do very lightweight operations there
if you have heavy processing either use synchronous application
and notify the asynchronous one passing it the results.

Or use the CppCMS's thread pool to handle heavy operations.

Don't forget that you can't access asynchronous app. directly
from other threads you rather need to use cppcms::service::post
to pass the handler to the main event loop for execution
to ensure thread safety.

Basically:

- When you need COMET/Server side events use asynchronous app,
- Otherwise use synchronous one.

Artyom


      

Re: [Cppcms-users] Mounting both synchronous and asynchronous applications

[CN <cn...@gr...>]

Many thanks for the quick help!

> you should provide a mount point
> that would distinguish between them:
>
> See:
>
> -
> http://art-blog.no-ip.info/cppcms_ref_v0_99/classcppcms_1_1applications__pool.html
>
> -  
> http://art-blog.no-ip.info/cppcms_ref_v0_99/classcppcms_1_1mount__point.html
>

Now I am trying to also provide mount() parameter mount_point but have  
trouble with the matching:

//code begin
class sync_app:public cppcms::application{ .... }
class async_app:public cppcms::application
{
	async_app(cppcms::service &srv) :	cppcms::application(srv)
	{
		dispatcher().assign("(.*)",&async_app::catch_all,this,1);
	}
	void catch_all(std::string url)
	{
		BOOSTER_LOG(debug,"URL") << url;
	}
}

int main(int argc,char **argv)
{
	cppcms::service s(argc,argv);
	s.applications_pool().mount(cppcms::applications_factory<sync_app>(),cppcms::mount_point("/s/a"));
	booster::intrusive_ptr<async_app> as_app=new interaction(s);
	//Watch the match string in the next line!
	s.applications_pool().mount(as_app,cppcms::mount_point("/as/a"));
	s.run();
}
//code end

Browser accesses to URL "/as/a". The match strings and logs follows:

match
string	log
---------------
/as	(none)
.*	/as/a
as	(none)
a	(none)
/a	(none)
/as/a	(none)

That being said, only provided with script ".*" will mount_point() fire  
application "async_app":

s.applications_pool().mount(i,cppcms::mount_point(".*"));

Any idea what mistake I have made?

Regards,
CN

Re: [Cppcms-users] Mounting both synchronous and asynchronous applications

[Artyom <art...@ya...>]

> >  -  
> >  
http://art-blog.no-ip.info/cppcms_ref_v0_99/classcppcms_1_1mount__point.html
> >
> 
> Now  I am trying to also provide mount() parameter mount_point but have  
> trouble with the matching:

First of all read once again this:
 http://art-blog.no-ip.info/cppcms_ref_v0_99/classcppcms_1_1mount__point.html

Most likely you are looking for:

   
http://art-blog.no-ip.info/cppcms_ref_v0_99/classcppcms_1_1mount__point.html#8aed02e81a2442db463910d47ac8d27a


Basically you are probably looking for

    s.applications_pool().mount(cppcms::applications_factory<sync_app>(),
              cppcms::mount_point("/s/(.*)",1));

     s.applications_pool().mount(as_app,
              cppcms::mount_point("/as/(.*)",1));

When you mount with one parameter you are using SCRIPT_NAME variable and 
PATH_INFO
passed to the main(url)

When you mount with string parameter and integer parameter you are using
PATH_INFO and match against url and then pass matched part to main

I don't know what is your web server configuration (or script names)
so I can't tell you what happens, best is to match against PATH_INFO




Artyom



      

Re: [Cppcms-users] Mounting both synchronous and asynchronous applications

[CN <cn...@gr...>]

My brain starts to malfunction due to late hour here. Sorry for the  
previous incorrect code. Please use this one instead. The unexpected  
results mentioned in last post remain unresolved.

//code begin
class sync_app:public cppcms::application{ .... }
class async_app:public cppcms::application
{
	async_app(cppcms::service &srv) :	cppcms::application(srv)
	{
		dispatcher().assign("(.*)",&async_app::catch_all,this,1);
	}
	void catch_all(std::string url)
	{
		BOOSTER_LOG(debug,"URL") << url;
	}
}

int main(int argc,char **argv)
{
	cppcms::service s(argc,argv);
	s.applications_pool().mount(cppcms::applications_factory<sync_app>(),cppcms::mount_point("/s/a"));
	booster::intrusive_ptr<async_app> as_app=new async_app(s); //typo error
	//Watch the match string in the next line!
	s.applications_pool().mount(as_app,cppcms::mount_point("/as/a"));
	s.run();
}
//code end

Browser accesses to URL "/as/a". The match strings and logs follows:

match
string	log
---------------
/as	(none)
.*	/as/a
as	(none)
a	(none)
/a	(none)
/as/a	(none)

That being said, only provided with script ".*" will mount_point() fire
application "async_app":

s.applications_pool().mount(as_app,cppcms::mount_point(".*"));

Regards,
CN

Re: [Cppcms-users] Mounting both synchronous and asynchronous applications

[CN <cn...@gr...>]

This morning this program all a sudden begins to match requested URI. The  
possible reason for being not working last night most likely is because  
the browser did not fully reload java script and/or the web page from web  
server (because of its caching mechanism?) even I clicked its refresh  
button. I had this experience before and wasted hours looking for the  
cause of unexpected behavior.

If people start to get the similar weird result, please try closing the  
tab from your browsers and then locate the URI again, or even restart it.


Back to my point... I am able to proceed my work now. However, there is  
still one thing I don't understand:

class async_app : public cppcms::application {
public:
	async_app(cppcms::service &srv) : cppcms::application(srv)
	{
		dispatcher().assign("^ps$",&async_app::catch_all,this);
	}
	void catch_all()
	{
		BOOSTER_LOG(debug,"async app is fired");
		std::map<std::string,std::string> env=request().getenv();
		std::map<std::string,std::string>::const_iterator i;
		for(i=env.begin();i != env.end();++i)
			BOOSTER_LOG(debug,"ENV") << (*i).first << ":" << (*i).second;
	}
}

int main(int argc,char **argv)
{
	cppcms::service s(argc,argv);
	s.applications_pool().mount(cppcms::applications_factory<sync_app>(),cppcms::mount_point("/s/(.*)",1));
	booster::intrusive_ptr<async_app> e=new async_app(s);
	s.applications_pool().mount(e,cppcms::mount_point("/as/(.*)",1));
	s.run();
	return 0;
}

The excerpted log follows:

2011-02-02 02:30:07 GMT; ENV, debug: DOCUMENT_ROOT:/var/www/example  
(cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: GATEWAY_INTERFACE:CGI/1.1  
(cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: HTTP_ACCEPT:text/html, */*  
(cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: HTTP_ACCEPT_CHARSET:iso-8859-1,  
utf-8, utf-16, *;q=0.1 (cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: HTTP_ACCEPT_ENCODING:deflate, gzip,  
x-gzip, identity, *;q=0 (cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: HTTP_CONNECTION:Keep-Alive, TE  
(cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: HTTP_HOST:www.example.com  
(cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug:  
HTTP_REFERER:http://www.example.com/test.html (cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: HTTP_TE:deflate, gzip, chunked,  
identity, trailers (cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: HTTP_USER_AGENT:Opera/9.80 (X11;  
Linux i686; U; en) Presto/2.7.62 Version/11.00 (cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: HTTP_X_REQUESTED_WITH:XMLHttpRequest  
(cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: PATH_INFO:/as/ps (cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug:  
PATH_TRANSLATED:/var/www/example/as/ps (cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: QUERY_STRING: (cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: REDIRECT_STATUS:200 (cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: REMOTE_ADDR:127.0.0.1 (cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: REMOTE_PORT:41362 (cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: REQUEST_METHOD:GET (cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: REQUEST_URI:/cgi/as/ps (cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: SCRIPT_FILENAME:/var/www/example/cgi  
(cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: SCRIPT_NAME:/cgi (cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: SERVER_ADDR:127.0.0.1 (cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: SERVER_NAME:www.example.com  
(cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: SERVER_PORT:80 (cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: SERVER_PROTOCOL:HTTP/1.1 (cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug:  
SERVER_SOFTWARE:lighttpd/1.4.28-devel-485M (cgi.cpp:114)

Please take special note of these two lines in the log:

2011-02-02 02:30:07 GMT; ENV, debug: PATH_INFO:/as/ps (cgi.cpp:114)
2011-02-02 02:30:07 GMT; ENV, debug: SCRIPT_NAME:/cgi (cgi.cpp:114)

The description for the overloaded function

cppcms::mount_point::mount_point(std::string const &script) 	

is

"Create a mount point that checks SCRIPT_NAME, and passes PATH_INFO for  
dispatching the version of mount_point() among the overloads"

Refer to:

http://cppcms.sourceforge.net/cppcms_ref_v0_99/classcppcms_1_1mount__point.html

My understanding is that the following code should match nothing:

	s.applications_pool().mount(e,cppcms::mount_point("/as/(.*)",1));

because according to the log, the incoming SCRIPT_NAME is "/cgi". However,  
the reality is that it matches! On the contrary, this version fails to  
match:

	s.applications_pool().mount(e,cppcms::mount_point("/cgi/(.*)",1));

Is this a descrepancy between CppCMS documentation and the library itself?

Best Regards,
CN

Re: [Cppcms-users] Mounting both synchronous and asynchronous applications

[Artyom <art...@ya...>]

> 
> Back  to my point... I am able to proceed my work now. However, there is  
> still one thing I don't understand:
> 
>      {
>          dispatcher().assign("^ps$",&async_app::catch_all,this);
>      }
>      
>s.applications_pool().mount(cppcms::applications_factory<sync_app>(),cppcms::mount_point("/s/(.*)",1));
>
>      s.applications_pool().mount(e,cppcms::mount_point("/as/(.*)",1));

Lets start:

Quoting Documentation:

> cppcms::mount_point::mount_point(std::string const & path,int group) 
>
> Create a mount point that checks PATH_INFO only and passes matched group for 
>dispatching 
>

That means that PATH_INFO is matched so 

   PATH_INFO = /as/ps

Matched agains:

   /as/(.*)

And then "ps" is selected and
passed to main of application
that on "^ps$" gets matched.

What does not match the documentation?

> cppcms::mount_point::mount_point(std::string const &script)
>
> Create a mount point that checks SCRIPT_NAME, and passes PATH_INFO for 
>dispatching 
>

Now this is fo case where we have for exaple

   foo.cgi
   bar.cgi

And you want to match them 

So for URL foo.cgi/beep SCRIPT_NAME=foo.cgi PATH_INFO=/beep

foo.cgi would be matched by mount_point("foo.cgi")
and then "/beep" would passed to main of the application for dispatching

...
> cppcms::mount_point::mount_point()
> 
> Create default mount point, it uses PATH_INFO for url-dispatching and gives no 
>restriction on URL 
>

This case - the default mount point when you use just a one application
it passes all PATH_INFO for URL dispatching.

> Please take special note of these two lines in the  log:
> 
> 2011-02-02 02:30:07 GMT; ENV, debug: PATH_INFO:/as/ps  (cgi.cpp:114)
> 2011-02-02 02:30:07 GMT; ENV, debug: SCRIPT_NAME:/cgi  (cgi.cpp:114)
> 
> The description for the overloaded  function
> 
> cppcms::mount_point::mount_point(std::string const &script)      
> 
> is
> 
> "Create a mount point that checks  SCRIPT_NAME, and passes PATH_INFO for  
> dispatching the version of  mount_point() among the overloads"
> 
> Refer  to:
> 
> 
http://cppcms.sourceforge.net/cppcms_ref_v0_99/classcppcms_1_1mount__point.html
> 
> My  understanding is that the following code should match  nothing:
> 
>      s.applications_pool().mount(e,cppcms::mount_point("/as/(.*)",1));
> 

No. It is matched against PATH_INFO that is the usual case

> because  according to the log, the incoming SCRIPT_NAME is "/cgi". However,  
> the  reality is that it matches! On the contrary, this version fails to  
> match:
> 
>      s.applications_pool().mount(e,cppcms::mount_point("/cgi/(.*)",1));
> 
> Is  this a descrepancy between CppCMS documentation and the library  itself?

No, you just should read documentation more carefully.

Different mount points constructors have different semantics allowing
to both use script_name and path_info for dispatching, however
every mount point should extract something to pass to application's main.

Artyom


      

Re: [Cppcms-users] Mounting both synchronous and asynchronous applications

[CN <cn...@gr...>]

Indeed, I appeared to be comparing apples and oranges.
Perhaps it's time for me to take some rest before I proceed.

Thank you for your time!

Regards,
CN