Martin Foster | 1 Sep 15:57 2006
Picon

Re: perl and "Handler cgi" under Windows

Alvaro Lopez Ortega <alvaro <at> gnu.org> writes:

>   Try this out:
> 
> =======
>   Extension pl {
>       Handler cgi {
> 	   ScriptAlias c:/perl/perl.exe
>       }
>   }
> =======
> 
>   I haven't tried it, but it should work.
>   Good luck!

Thanks for your quick reply,

Unfortunately, I just get a nondescript 500 error, with nothing explanatory on
the console.

I tried replacing the path to perl.exe with the path to the compiled version of:

======
#include <stdio.h>

int main (int argc, char **argv) {
	puts("Content-type: text/plain;\n");
	int i;
	for (i = 0; i < argc; ++ i) printf("[%s]\n", argv[i]);
	return 0;
(Continue reading)

JJL | 4 Sep 16:11 2006
Picon

trac multi projects and fcgi

Hi,

For my first post, I would like to Thanks you for such a pretty web server.
BTW, here's my problem :

I'm trying to use trac for many projects in a FastCGI environement.
I finaly succedded after many tries and many list-search
But I have a cosmetic problem.

Here's my configuration :
Server trac.mydomain.com {
   DocumentRoot /usr/share/trac/cgi-bin
   DirectoryIndex trac.fcgi
   Directory / {
      Handler common
   }
   Request "^/.+/chrome/common/(.*)$" {
      Handler redir {
         Show Rewrite "/chrome/common/$1"
      }
   }
   Directory /chrome/common {
      Handler file
      DocumentRoot /usr/share/trac/htdocs/
    }
   Extension fcgi {
      Handler fcgi {
         ErrorHandler on
         CheckFile off
         Server localhost:8001 {
            Env TRAC_ENV_PARENT_DIR "/var/trac"
            Interpreter "/usr/share/trac/cgi-bin/trac.fcgi"
         }
      }
   }
}
So, when I acess http://trac.mydomain.com/ I have a list of my projects and I can access to every trac functions for each of them.
But url are http://trac.mydomain.com/trac.fcgi/project1, http://trac.mydomain.com/trac.fcgi/project2/browser ...
And I find this 'trac.fcgi' quite ugly in the url :( I would prefer some http://trac.mydomain.com/project2/browser

Is there a way to get rid of it ?

Thanks
JJL
PS: I'm using Cherokee 0.5.4 and trac 0.9.3

--
http://kubuntu.free.fr

_______________________________________________
Cherokee mailing list
Cherokee <at> 0x50.org
http://www.0x50.org/cgi-bin/mailman/listinfo/cherokee
Alvaro Lopez Ortega | 4 Sep 16:40 2006
Picon

Re: trac multi projects and fcgi

JJL wrote:

> So, when I acess http://trac.mydomain.com/ I have a list of my
> projects and I can access to every trac functions for each of them.
> But url are http://trac.mydomain.com/trac.fcgi/project1,
> http://trac.mydomain.com/trac.fcgi/project2/browser
> <http://trac.mydomain.com/trac.fcgi/project2/browser> ...  And I
> find this 'trac.fcgi' quite ugly in the url :( I would prefer some
> http://trac.mydomain.com/project2/browser
>
> Is there a way to get rid of it ?

  The easiest way is to add a Request entry to your configuration
  file, although it's more a work around than a proper solution.

  This should be enough (add it after the "Extension fcgi" entry):

=======
  Request "^/trac.fcgi/(.*)$" {
    Handler redir {
      Rewrite "/$1"
    }
  }
=======

--

-- 
Greetings, alo.
http://www.alobbs.com
Antonio Galindo Castro | 4 Sep 21:37 2006
Picon

Tiny Howto install under Solaris 8.

Today I have Cherokee 0.5.4 running over a Solaris 8.  This is mi litle Howto.

###  The machine.
    bash-2.03# uname -a
    SunOS pruebas 5.8 Generic_117350-05 sun4u sparc SUNW,UltraAX-i2

    ###  Optimization for the machine.
    export CXXFLAGS='-O2 -mcpu=ultrasparc'
    export CFLAGS='-O2 -mcpu=ultrasparc'

###  If you have problems to found cru (ar), this is for you.
    export PATH=$PATH:/usr/ccs/bin/

###  Disabling "problematic" things (In Solaris 8)...
    ./configure --prefix=/usr/local --disable-tls --disable-pam

###  Fixating problems whith readdir_r
    vi +409 cherokee/util.c
    :407s/^/\/*/
    :417s/$/*\//

    adding one line (419):
    return readdir_r (dirstream, entry, result);

###  Another solution for the problems whith readdir_r.

    vi +409 config.h
    original:
    #define HAVE_READDIR_R 1
    before:
    /*#define HAVE_READDIR_R 1*/

    line 912:
    original:    #define HAVE_READDIR_R_2 1
    before:
    /*#define HAVE_READDIR_R_2 1*/

    line 915:
    original:
    /* #undef HAVE_READDIR_R_3 */
    before:
    #define HAVE_READDIR_R_3 1

###  Fixating the problem to found getopt.h
    cp cherokee/getopt/getopt.h cherokee/
    cp cherokee/getopt/getopt.h cget/

###  Finaly:
    make
    make check
    make install

###  Enjoy cherokee!  :D

 Cheers... 

_______________________________________________
Cherokee mailing list
Cherokee <at> 0x50.org
http://www.0x50.org/cgi-bin/mailman/listinfo/cherokee
Manuel Arostegui Ramirez | 4 Sep 22:06 2006

Re: Tiny Howto install under Solaris 8.

El Lunes, 4 de Septiembre de 2006 21:37, Antonio Galindo Castro escribió:
> Today I have Cherokee 0.5.4 running over a Solaris 8.  This is mi litle
> Howto.
>
> ###  The machine.
>     bash-2.03# uname -a
>     SunOS pruebas 5.8 Generic_117350-05 sun4u sparc SUNW,UltraAX-i2
>
>     ###  Optimization for the machine.
>     export CXXFLAGS='-O2 -mcpu=ultrasparc'
>     export CFLAGS='-O2 -mcpu=ultrasparc'
>
> ###  If you have problems to found cru (ar), this is for you.
>     export PATH=$PATH:/usr/ccs/bin/
>
> ###  Disabling "problematic" things (In Solaris 8)...
>     ./configure --prefix=/usr/local --disable-tls --disable-pam
>
> ###  Fixating problems whith readdir_r
>     vi +409 cherokee/util.c
>
>     :407s/^/\/*/
>     :417s/$/*\//
>
>     adding one line (419):
>     return readdir_r (dirstream, entry, result);
>
> ###  Another solution for the problems whith readdir_r.
>
>     vi +409 config.h
>     original:
>     #define HAVE_READDIR_R 1
>     before:
>     /*#define HAVE_READDIR_R 1*/
>
>     line 912:
>     original:    #define HAVE_READDIR_R_2 1
>     before:
>     /*#define HAVE_READDIR_R_2 1*/
>
>     line 915:
>     original:
>     /* #undef HAVE_READDIR_R_3 */
>     before:
>     #define HAVE_READDIR_R_3 1
>
> ###  Fixating the problem to found getopt.h
>     cp cherokee/getopt/getopt.h cherokee/
>     cp cherokee/getopt/getopt.h cget/
>
> ###  Finaly:
>     make
>     make check
>     make install
>
> ###  Enjoy cherokee!  :D
>
>  Cheers...

Great!
Could be a great idea to upload it to the cherokee wiki :-)
http://www.0x50.org/wiki/index.php?title=Main_Page

Kind regads.

--

-- 
Manuel Arostegui Ramirez.

Electronic Mail is not secure, may not be read every day, and should not
be used for urgent or sensitive issues.
Alvaro Lopez Ortega | 4 Sep 22:28 2006
Picon

Re: Tiny Howto install under Solaris 8.

Manuel Arostegui Ramirez wrote:

>> Today I have Cherokee 0.5.4 running over a Solaris 8.  This is mi
>> litle Howto.

  Good stuff Antonio! :-)

> Could be a great idea to upload it to the cherokee wiki :-)
> http://www.0x50.org/wiki/index.php?title=Main_Page

  Even if it usually is a good idea, I don't think we should do it in
  this case.  If the compilation isn't straight forward on Solaris 8,
  we should consider it as a bug that we have to fix rather than start
  documenting the work around.

  Actually, Antonio's how to is going to be very helpful to allow us
  to fix the issues in the next release.

--

-- 
Greetings, alo.
http://www.alobbs.com
Jordi Adame | 5 Sep 20:12 2006
Picon

PHP and cherokee

hello people,

I've been working a high load site with cherokee and PHP, the only issue i've had is that sometimes pages appear incomplete, as if cherokee had died in the middle of the request.

I've compiled php4 and 5 several times and i haven't got any better results, I thought it was a PHP problem, but after doing some tests with apache and php-fcgi i got no incomplete pages with this conf.

I was wondering if there is any way to get a debug cherokee, or anything that would help me figure this out.

cheers
jordi

_______________________________________________
Cherokee mailing list
Cherokee <at> 0x50.org
http://www.0x50.org/cgi-bin/mailman/listinfo/cherokee
Alvaro Lopez Ortega | 5 Sep 20:34 2006
Picon

Re: PHP and cherokee

Jordi Adame wrote:

> I've been working a high load site with cherokee and PHP, the only
> issue i've had is that sometimes pages appear incomplete, as if
> cherokee had died in the middle of the request.
>
> I've compiled php4 and 5 several times and i haven't got any better
> results, I thought it was a PHP problem, but after doing some tests
> with apache and php-fcgi i got no incomplete pages with this conf.
>
> I was wondering if there is any way to get a debug cherokee, or
> anything that would help me figure this out.

  Add the parameter --enable-trace when you execute the configure
  script;  It'll enable the tracing facility.

  Then, you will be able to inspect what the server is doing by using
  the CHEROKEE_TRACE environment variable. For example:

      CHEROKEE_TRACE="all" /usr/sbin/cherokee

  That will print every single step that the server performs. In this
  case I suppose you don't need so much information, so something like
  CHEROKEE_TRACE="fcgi" or CHEROKEE_TRACE="handler" would be enough.

  By the way, this tracing mechanism *slightly* reduces the server
  performance, but it isn't very much actually, so don't worry about
  it.

  Oh! And finally, if you check the source code and see places in
  which it could be good to put new TRACE entries, do not hesitate to
  submit the patch.

--

-- 
Greetings, alo.
http://www.alobbs.com
Alvaro Lopez Ortega | 5 Sep 22:54 2006
Picon

New MySQL validator

Hi folks,

  I just wanted to announce that a brand new MySQL validator has been
  imported into the unstable branch (the upcoming 0.6 release).

    http://www.0x50.org/bugs/browser/cherokee/trunk/cherokee/validator_mysql.c

  Brian Rosner has done a really good job writing this first version.
  (Thanks man, you rock! :-)

  As usual, feedback is welcome.. so, keep the ideas coming.

--

-- 
Greetings, alo.
http://www.alobbs.com
Eric S. Johansson | 6 Sep 05:43 2006

301 puzzlement

I am probably missing something obvious again but here's something that 
has me puzzled.  I have a CGI program that executes for every URL 
starting with /akasha.  But if I leave off the trailing slash, I get a 
redirect to the URL with a /.  Ideally what should happen is it 
shouldn't matter if there's a trailing slash, it should just invoke the 
CGI and be done with it.  normally this is not a problem when accessing 
pages but it is causing problems for CGI.  It would really be nice to 
not have to remember to add that trailing /

what am I missing?

Directory /akasha {

     Handler cgi {
         ErrorHandler on
         scriptalias /var/www/scripts/default/ak_cgi.cgi

         # point to the akasha modules
         Env PYTHONPATH "/usr/local/python/akasha"

         # point to the default css file
         Env AKASHA_CSS "/files/akasha.css"

         # base_path is the full pathname to the akasha data store
         # don't forget the trailing slash
         # note: this path should not be exposed to the outside world
         Env AKASHA_BASE_PATH "/home/esj/projects/data-akasha/"

         # make this the same as the script name in scriptalias
         Env AKASHA_SCRIPT "ak_cgi.cgi"
     }
}

*** ethereal capture.

GET /akasha HTTP/1.1
Host: 192.168.118.16
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.0.6) 
Gecko/20060728 Firefox/1.5.0.6
Accept: 
text/xml,application/xml,application/xhtml+xml,text/html;q=0.9,text/plain;q=0.8,image/png,*/*;q=0.5
Accept-Language: en-us,en;q=0.5
Accept-Encoding: gzip,deflate
Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.7
Keep-Alive: 300
Connection: keep-alive

HTTP/1.1 301 Moved Permanently
Connection: Close
Date: Tue, 05 Sep 2006 23:31:20 GMT--5
Server: Cherokee/0.5.4 (UNIX)
Location: /akasha/
Content-Type: text/html
Content-length: 267
Cache-Control: no-cache
Pragma: no-cache
P3P: CP=3DNOI NID CURa OUR NOR UNI

Gmane