Re: possible issue in mod_perl 2.0.4 with respect to STDERR
On Thu 12 Nov 2009, Shibi NS wrote: Configuration error log is in my applications httpd config file is ErrorLog = '| rotatelogs \log\error_log 1M ' But if print something on STDERR from mod_perl program , Say debug print (print STDERR DEBUG ;) this is going to \log\error_log instead of \log\error_log.NNN Is this is a mod perl issue or i am doing something wrong I suspect this ErrorLog directive resides inside a VHost? STDERR is bound to the programs notion of file descriptor 2 or whatever it is on windows, not to the ErrorLog. If you want to write to a VHost's ErrorLog use $r-log co. See also http://www.gossamer-threads.com/lists/modperl/modperl/99241 http://www.gossamer-threads.com/lists/modperl/modperl/99219 http://www.gossamer-threads.com/lists/modperl/modperl/99504 In addition to what is said here http://www.gossamer-threads.com/lists/modperl/modperl/99248#99248 another way to capture STDERR could be a PERLIO layer or to tie STDERR to something. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: Storing config values in-memory between sessions
On Thu 29 Oct 2009, Mahesh Khambadkone wrote: As it seldom changes, we dont want to use a database for these 'config values', yet need a way to retain in memory and dirty its' value from time to time. Have a look at MMapDB which I have just uploaded to CPAN. I wrote this module some time ago exactly for that case: configuration data that seldom changes. Before I have used BerkeleyDB. But that is a bit touchy when it comes to sudden process death. I wanted something simple, stable and fast. So an MMapDB database cannot be modified by a connected process because it is mapped read-only. Changing data means rewriting the database file and reconnect the new one. This also means the whole database is shared memory between all connected processes. There is a tie() based interface. So, even dumping human readable data is simple: perl -MMMapDB -MData::Dumper -e ' print Dumper(MMapDB-new(filename=shift)-start-main_index); ' /etc/opt/TRANSCONFIG/transconfig.mmdb And a more complicated but faster interface: perl -MMMapDB -le ' # open the DB my $db=MMapDB-new(filename=shift)-start; # get data record positions my @pos=$db-index_lookup($db-mainidx, qw!actn opi /svn!); # fetch a data item print $db-data_record($pos[0])-[2]; ' /etc/opt/TRANSCONFIG/transconfig.mmdb One even can store large values there and pass references to them around and thus avoid copying and mallocing that happens normally when you do $x=$y: perl -MMMapDB -MDevel::Peek -e ' my $db=MMapDB-new(filename=shift)-start; my @pos=$db-index_lookup($db-mainidx, qw!actn opi /svn!); # get a reference to the data item to avoid copying my $v=\$db-data_record($pos[0])-[2]; print $$v, \n; Dump $$v ' /etc/opt/TRANSCONFIG/transconfig.mmdb File: '/opt/svnbook'.$MATCHED_PATH_INFO SV = PV(0x9d4b50) at 0x7c9248 REFCNT = 1 FLAGS = (POK,READONLY,pPOK) PV = 0x7f956388d240 File: '/opt/svnbook'.$MATCHED_PATH_INFO CUR = 39 LEN = 0 You see $v points to a read-only variable. The PV pointer of this variable references the string inside the read-only mapped database file. Now you can do $x=$v and only the reference is copied. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: Server hangs after 502 request
On Sat 24 Oct 2009, Rolf Schaufelberger wrote: Starting some weeks ago the server sometimes hangs with an out of memory problem. Assuming you are running on linux the following sysctls may help to find the culprit. vm.overcommit_memory=2 vm.overcommit_ratio=90 By default (overcommit_memory=0) when a process needs more memory linux simply says okay here you are no matter if the memory is currently available or not. This is based on the assumption that most processes do not use all memory they allocate at all. Later on, when the process really accesses the memory a page fault is generated and linux only then allocates the memory for the process. But now it is too late to signal the process out-of-memory. So, linux somehow has to obtain memory no matter what. So, in short-of-memory conditions linux starts the OOM killer. It implements some heuristic that chooses some best fitting processes to be killed. These best fitting processes may be totally unrelated to the original problem. I had once a case where a perl program processed mailbox files (using Mail::Box) on a box where a postgres database ran. Unfortunately Mail::Box reads in the whole mailbox file. Normally our mailbox files were about 1-10mb and the program had worked for years. But suddenly we had one of 1Gb. Instead of signaling out-of-memory to the perl process linux killed the postgres database. Make sure you have enough swap space (at least the RAM size) before experimenting with those sysctls. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: Custom INC per-directory
On Wed 21 Oct 2009, Alan Young wrote: http://perl.apache.org/docs/2.0/user/config/config.html#C_Parent_ Unfortunately the Parent option works only in vhosts. But I believe it could be hacked to work also for directory containers. Of course PerlPostReadRequest, PerlTranslation and PerlMapToStorage phases are handled by the main interpreter for the vhost. But just before the PerlHeaderParserHandler is called one could check if the rest of the request cycle is to be handled by a different interpreter and fetch one from a cache or instantiate a new one. Of course the user would have to know what he is doing using this feature. For example setting a PerlResponseHandler where the handler routine is an anonymous sub in the translation phase would not work as expected. An alternative would be to have modperl inspect the uri prior to PerlPostReadRequest and decide what interpreter to use based on it, similar to what ServerPath does. Then the same interpreter would handle all phases. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: is it possible get the deflate information in reuest object
On Mon 12 Oct 2009, Shibi NS wrote:
LogFormat '%r %{outstream}n/%{instream}n (%{ratio}n%%)
%{User-agent}i' deflate
CustomLog /var/log/httpd/deflate_log deflate
Can I get the Input/Output and Ratio information in Apache2 Request
object or PerlLogHandler ?
in a PerlLogHandler you should be able to read them as
$r-notes-{ratio} etc.
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
Re: Configuring virtual hosts on the fly
On Sun 11 Oct 2009, André Warnier wrote: I see some hooks in PerlTransHandler and PerlMapToStorageHandler that seem like they can almost do what I want, but I don't see how to set other virtual host parameters, like ServerAdmin http://perl.apache.org/docs/2.0/api/Apache2/ServerRec.html#C_server_a dmin_ , UseCanonicalName, etc. I'm just starting to work on this, and I thought I would see if anybody had tried anything similar before and had some suggestions for what to do, or what not to do. I have never tried this myself, but the mod_perl 2.x documentation contains some hints that leads to believe that it should be possible. Some things could work this way, others definitely will not. For example you cannot create an IP based vhost on demand. You cannot create log files on the fly. Well, you can but you cannot create them as root as it is done during startup. In general its not a good idea to mess with things that are not marked as applicable in directory context for the following reason. Firstly, you have to make sure to reset them when you request is done. See the implementation of $r-document_root for example. And secondly, in a threaded environment changing those values in one thread can affect the behavior of other requests in other threads because they are shared. UseCanonicalName is applyable in directory context. So, you can safely set it via $r-add_config. It may however be useless to set it in a translation or map2storage handler because if you have a directory or location container or even a .htaccess file that sets the directive it overwrites anything set in a translation handler. Best phase to do that is probably fixup. ServerAdmin is valid only in server config and virtual host context (http://httpd.apache.org/docs/2.2/mod/core.html#serveradmin). So I wouldn't mess with that. Better create custom error pages. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: Threading inside a handler script
On Sat 03 Oct 2009, Elizabeth Mattijsen wrote: I believe what you're doing should work, but I don't use threads for anything. I'd probably fork and put the results in a database table or file. If you like the threads API, you might want to take a look at the forks module on CPAN. It provides the threads API using fork(), and hence does not have many of the problems that native Perl ithreads have. Or Coro. I have never used it with modperl but I think it will work. It does not use kernel threads but implements perl stack switches. So, you have to watch out to avoid blocking operations. Similar to win 3.x multitasking or early UNIX threading approaches. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: user undef
On Wed 30 Sep 2009, Bernhard Janetzki wrote:
Hello together,
i'm trying basic authentication via
## apache2 config:
Location /auth
SetHandler perl-script
PerlAuthenHandler DataExchange::AuthHandler
#PerlResponseHandler ModPerl::Registry
Options +ExecCGI
PerlOptions +GlobalRequest
AuthType Basic
AuthName The Gate
Require valid-user
/Location
## Example script:
package DataExchange::AuthHandler;
use strict;
use warnings;
use Apache2::Access ();
use Apache2::RequestUtil ();
use Apache2::Const -compile = qw(OK HTTP_UNAUTHORIZED);
sub handler {
my $r = shift;
warn $r-user();
my $password = $r-get_basic_auth_pw;
have a look at
http://perl.apache.org/docs/2.0/api/Apache2/Access.html#C_get_basic_auth_pw_
get_basic_auth_pw returns a status and the password. Your code has to
check that status before proceeding.
see also
http://perl.apache.org/docs/2.0/user/handlers/http.html#PerlAuthenHandler
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
Re: ModPerl::Registry and Error Documents
On Tue 29 Sep 2009, Andreas Mock wrote: after searching for a while without finding an answer to my question I hope you can help me, We're using mod_perl with ModPerl::Registry to have only a thin layer between apache and our perl scripts. So far so good. Now we want to produce error documents dynamically and thought this could be achieved by simply producing a normal html document and setting a status code 400. But it seems that the apache error handling jumps in as soon as the code is seen and the html document is not delivered to the client. My question: Which is an easy and correct way to produce ErrorDocuments while using ModPerl::Registry? You have asked this question already 2 times before, on 17.09.09 12:13 and 17.09.09 11:15. The first time I have answered it. If my answer is not satisfying could you be so kind and explain why? If you want it to work the way you anticipate it should and it does not then take a fresh copy from the SVN repository at http://svn.eu.apache.org/repos/asf/perl/modperl/trunk and provide a patch. But please stop asking the same question again and again. Glücklicherweise spreche ich die selbe Sprache wie Du. Also, Du hast diese Frage schon 2 Mal gestellt. Beim ersten Mal habe ich geantwortet. Wenn Dir meine Antwort nicht gefällt, könntest Du bitte erklären, was Du daran auszusetzen hast? Wenn modperl nicht so tut, wie Du denkst, es sollte, dann nimm einen frischen Checkout von http://svn.eu.apache.org/repos/asf/perl/modperl/trunk bau es nach Deinem Geschmack um und stelle einen Patch zur Verfügung. Das ist der beste/einzige Weg, die gewünschte Funktionalität in modperl zu kriegen. Ich denke auch, daß es mit ParseHeaders und einer Status: NNN Zeile tun sollte. Solange ich aber nicht direkt vor dem Problem stehe oder nichts anderes zu tun habe, was nie passieren wird, werde ich das Problem nicht angehen. So ähnlich denken wohl auch die anderen Entwickler. Bitte hör auf, ein und die selbe Frage immer wieder zu stellen. Und übrigens, der Text in meiner Signatur ist ernst gemeint. Ich bin käuflich. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: Updating cookies in header during request processing
On Fri 18 Sep 2009, Igor Chudov wrote: But how can I change the cookie AFTER I called $cgi-start_html? As long as there is nothing sent out on the wire you can add headers. When the response header is gone you can't. start_html() only generates a string as far as know. So it depends upon when this string is sent. But with mod_perl you can circumvent even this limitation. Write a caching output filter. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: ModPerl::Registry and custom error documents
On Thu 17 Sep 2009, Andreas Mock wrote: How can we dynamically create own error documents without using the lower level mod_perl/apache api? Is there a simple way? How can we achieve that? A very simple registry script: #!/usr/bin/perl use strict; my $r=shift; $r-status($r-args); exit; Then you can configure ErrorDocuments, e.g. ErrorDocument 404 http://huhu.com ErrorDocument 500 dumm gloffe Now try it out: A 404 will be translated into a redirect: $ curl http://localhost/etest?404 -v * About to connect() to localhost port 80 (#0) * Trying ::1... connected * Connected to localhost (::1) port 80 (#0) GET /etest?404 HTTP/1.1 User-Agent: curl/7.19.0 (x86_64-suse-linux-gnu) libcurl/7.19.0 OpenSSL/0.9.8h zlib/1.2.3 libidn/1.10 Host: localhost Accept: */* HTTP/1.1 302 Found Date: Thu, 17 Sep 2009 11:05:19 GMT Server: Apache/2.2.13 (Unix) proxy_html/3.0.1 mod_ssl/2.2.13 OpenSSL/0.9.8h DAV/2 SVN/1.6.4 mod_apreq2-20090110/2.7.1 mod_perl/2.0.5threading3 Perl/v5.10.0 Location: http://huhu.com Content-Length: 199 Content-Type: text/html; charset=iso-8859-1 !DOCTYPE HTML PUBLIC -//IETF//DTD HTML 2.0//EN htmlhead title302 Found/title /headbody h1Found/h1 pThe document has moved a href=http://huhu.com;here/a./p /body/html * Connection #0 to host localhost left intact * Closing connection #0 a 403 sends the normal error page $ curl http://localhost/etest?403 -v * About to connect() to localhost port 80 (#0) * Trying ::1... connected * Connected to localhost (::1) port 80 (#0) GET /etest?403 HTTP/1.1 User-Agent: curl/7.19.0 (x86_64-suse-linux-gnu) libcurl/7.19.0 OpenSSL/0.9.8h zlib/1.2.3 libidn/1.10 Host: localhost Accept: */* HTTP/1.1 403 Forbidden Date: Thu, 17 Sep 2009 11:05:58 GMT Server: Apache/2.2.13 (Unix) proxy_html/3.0.1 mod_ssl/2.2.13 OpenSSL/0.9.8h DAV/2 SVN/1.6.4 mod_apreq2-20090110/2.7.1 mod_perl/2.0.5threading3 Perl/v5.10.0 Content-Length: 207 Content-Type: text/html; charset=iso-8859-1 !DOCTYPE HTML PUBLIC -//IETF//DTD HTML 2.0//EN htmlhead title403 Forbidden/title /headbody h1Forbidden/h1 pYou don't have permission to access /etest on this server./p /body/html * Connection #0 to host localhost left intact * Closing connection #0 and a 500 delivers a custom error message: $ curl http://localhost/etest?500 -v * About to connect() to localhost port 80 (#0) * Trying ::1... connected * Connected to localhost (::1) port 80 (#0) GET /etest?500 HTTP/1.1 User-Agent: curl/7.19.0 (x86_64-suse-linux-gnu) libcurl/7.19.0 OpenSSL/0.9.8h zlib/1.2.3 libidn/1.10 Host: localhost Accept: */* HTTP/1.1 500 Internal Server Error Date: Thu, 17 Sep 2009 11:09:19 GMT Server: Apache/2.2.13 (Unix) proxy_html/3.0.1 mod_ssl/2.2.13 OpenSSL/0.9.8h DAV/2 SVN/1.6.4 mod_apreq2-20090110/2.7.1 mod_perl/2.0.5threading3 Perl/v5.10.0 Content-Length: 11 Connection: close Content-Type: text/html; charset=iso-8859-1 * Closing connection #0 dumm gloffe Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: Why people not using mod_perl
On Thu 17 Sep 2009, Kiran Kumar wrote: There is also Padre (http://padre.perlide.org/) , You can write plugins and customize to your needs, there are already lots of plugins available http://search.cpan.org/search?query=padre%3A%3Apluginmode=all I have seen padre first time at the this year German perl workshop in February and tried it out a bit. What I miss is syntax highlighting and indentation for C, XS and Perl in one tool. Can padre handle this? Last time I looked it could not but that was half a year ago. I am using Emacs for almost 20 years now but it lacks good XS support. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: Ways to scale a mod_perl site
On Wed 16 Sep 2009, Igor Chudov wrote: I have very little static content. Even images are generated. My site generates images of math formulae such as (x-1)/(x+1) on the fly., I can understand generating them on the fly for flexibility reasons, but I'd cache them, and serve them statically after that, rather than regenerate the images on every single request. You can accomplish that in the app itself, or just by throwing a caching proxy in front of it (maybe you're already doing this with perlbal) I actually do cache generated pictures, I store them in a database table called 'bincache'. This way I do not have to compute and draw every image on the fly. If I have a picture in bincache, I serve it, and if I do not, I generate it and save it. That saves some CPU, but makes mysql work harder. I'd go for Apache's mod_cache + mod_disk_cache. The only thing you have to do is to set cache control headers. Mod_cache is really fast b/c it skips almost all of the http request cycle. And in your case it takes load from the database. The request won't even hit mod_perl. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: changing apache process name?
On Tue 01 Sep 2009, Sebastiaan Hoogeveen wrote: Not sure if this would fix your problem, but you can just set $0 from within the mod_perl module to whatever you'd like to show up in the process list, just as with a regular perl script. It probably depends on the OS you're running whether this works as it should; I can confirm that it works on Linux. This should work in mod_perl as well. Perl 5.8.0 had a bug that prevented it. Hence the Sys::Proctitle module. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: changing apache process name?
On Tue 01 Sep 2009, mcvella wrote: I am wondering if anyone knows if there is a way to configure what it shown as the running apache process name. I am not sure if this is technically a mod_perl question. The problem is that the process name is partially truncated, so we are unable to determine exactly what perl app is running for a given apache process. For example, here is one running apache process as seen in top: 31122 www 16 0 76624 61m 6428 S 6 0.4 0:04.03 /opt/myapp/www/cgi-perl/sig -f /opt/c2/preview/tmp/httpd.conf there is more to the perl app name than 'sig', (for example, it could be signal.pl or signoff.pl) but it is truncated. Is there any way to configure apache to show more characters in its process names? It depends upon your operating system. FreeBSD has a library function called setproctitle to do that as far as I know. On Linux it can be achieved like this. The original argv and envp are laid out subsequently by the kernel. Normally a program accesses its argv only at startup and the original environment is copied and not used later on. But the argv pointer is where ps and top look for the command line. So there is a buffer of the size of all command line arguments plus all the environment that can be overwritten and thus used to show what you like in ps and top. The only problem is how to get the values of these pointers. This can be solved using a shared library. It may contain a function named _init() that is called when the lib is loaded. Fortunately it is passed argc, argv and envp as parameters. Now, you have all information to figure out the size of the argv and envp buffer and can overwrite it as you like. This is what the libsetproctitle (http://lmgtfy.com/?q=libsetproctitle) library does. I have once (at the time of perl 5.8.0) written a module Sys::Proctitle that uses (and bundles) this lib and also an Apache2::ShowStatus that uses Sys::Proctitle to show the current request in top. But since then I haven't touched these modules. So I don't know if they still work. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: mod_perl2 interactive startup
On Tue 25 Aug 2009, ama...@solutionsforprogress.com wrote: Unfortunately using restart_count does not address my issue. In the http.conf I have the following: ... PerlConfigRequire conf/startup.pl ... My test startup.pl has this: our $INPUT; print STDERR Input:; $INPUT = STDIN; print STDERR you entered $INPUT; My issue is during the first start everything works as I need it to, Input: is printed out to the terminal and the script waits for input. The problem is for the automatic restart everything is tied to the log files. Everything printed to STDERR goes into the logs and a user cannot enter any input. For the restart how can I print messages and prompt for input? Don't know if it actually works, but how about POSIX::dup2()ing STDERR/STDIN (or better opening /dev/tty and use this one for reading and writing) to well known file descriptors (say 1000, 1001) when restart_count==0. Then when restart_count0 you should be able to use them. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: getting $r and env to reflect external ports
On Sat 15 Aug 2009, Jonathan Swartz wrote:
We've got a bunch of legacy code that uses $ENV{SCRIPT_URI},
$ENV{SERVER_PORT}, $r-server-port and the like to generate external
redirects.
This has worked fine til now, but now we are moving to a system where
our external port != our internal port - the load balancer is going
to forward ports 80 and ports 443 to non-secure ports 8080 and 8443
(so that we can run httpd as a non-root user.) Unfortunately, the
code mentioned above will then use the wrong (internal) ports in the
redirects. If I were using Apache reverse proxy, ProxyPassReverse
would take care of this, but I'm not.
The legacy code is scattered in many files and will be difficult to
change. Thus I'm wondering if there's an easy way to get %ENV and $r
to reflect the external ports, without going through and changing
each port-related $ENV and $r value by hand.
e.g. In the Apache code I see that the ENV vars are populated via
ap_get_server_port - is there a way for me to hook in early enough
that I can change this?
According to the docs the server port can also be set. But I think this is not
a good idea.
The $r-server-port part can possibly be circumvented by reassigning a new
function to the Apache2::ServerRec::port symbol at startup time. Try this in
your startup.pl:
use Apache2::ServerRec ();
BEGIN {
my $old=\Apache2::ServerRec::port;
*Apache2::ServerRec::port=sub {
if( @_==1 ) {# read access
my $port=$old-(@_);
if( $port==8080 ) {
return 80;
} elsif(...) {
...
} else {
return $port;
}
} else {
return $old-(@_);
}
};
}
As for the %ENV part this approach may work. Call $r-subprocess_env in void
context without parameters somewhere before the response phase, e.g. fixup.
This initializes the CGI variables in %ENV or rather $r-subprocess_env. Then
use $r-subprocess_env to change what you need. The subsequent perl-script
handler will know that the CGI variables are already set up and won't do it
again. So your script should see the overridden values. That's the theory.
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
Re: disable input and output filters for subfolders
On Sat 01 Aug 2009, Mike Barborak wrote:
Invalid per-directory PerlOption: InputFilter
so it seems that while PerlOption is allowed at the directory level,
the InputFilter and OutputFilter options are not. (i actually
couldn't find a relevant section in the docs about that so i am
unsure.)
This might quite be true.
it is my preferred solution to be able to turn the filter off for
particular directories via the conf file so it is my fading hope that
i'm doing something wrong?
if i am not then my plan of attack is to use PerlSetVar. then my conf
file would look like so:
Directory /var/www/vhosts/mydomain.com/httpdocs/test
PerlSetVar EnableMyFilter false
/Directory
and my input and output filter handlers would look something like
this:
sub handler : FilterRequestHandler
{
my $f = shift;
return Apache2::Const::DECLINED if $f-r-dir_config (
'EnableMyFilter' ) eq 'false';
...
}
does that approach sound correct?
Almost, it would be better to have the filter remove itself on first
invocation if it's not needed. A typical filter of mine looks like
this:
my $ctx=$f-ctx;
unless( defined $ctx ) {
unless( check_conditions($f) ) {
$f-remove;
return Apache2::Const::DECLINED;
}
$f-ctx($ctx=[]);
$f-r-headers_out-unset('Content-Length');
}
...
$f-ctx is undefined on first invocation. So, unless defined $ctx I can
run checks that need to be done only once. If one of the conditions is
not met the filter removes itself and returns DECLINED.
You'll have to place your PerlVar check instead of check_conditions.
or is that i need to add an enabling / disabling filter into the
filter chain in front of my filter that reads the PerlSetVar option
and then either leaves my filter in the chain or uses something like
$f-next-next...-remove to remove it?
No, this is only a last resort if you cannot change the filter itself.
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
Re: disable input and output filters for subfolders
On Wed 29 Jul 2009, Mike Barborak wrote: i have filters working correctly for a directory on my website. there is a subdirectory that i would like to not be passed through the filters. can i do this through the apache conf file or do i need to code this into my filters? here is my apache setup: PerlRequire /var/www/vhosts/mydomain.com/conf/seo_startup.pl PerlModule HTTPRequestFilter PerlModule HTMLFilter Directory /var/www/vhosts/mydomain.com/httpdocs PerlInputFilterHandler HTTPRequestFilter PerlOutputFilterHandler HTMLFilter /Directory grasping for straws, i tried to add this but it didn't work: Directory /var/www/vhosts/mydomain.com/httpdocs/test PerlInputFilterHandler PerlOutputFilterHandler /Directory is there magic to remove filters for a subdirectory or even disable mod_perl altogether? A filter can remove itself on first invocation ($f-remove). It can also remove any filter further down the filter chain ($f-next-next...-remove). Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: disable input and output filters for subfolders
On Thu 30 Jul 2009, Torsten Foertsch wrote: Sorry, I hit return while one of my left hand fingers rested on the CTRL key and the MUA interpreted this as send mail now. So, here comes the rest of what I wanted to say. On Wed 29 Jul 2009, Mike Barborak wrote: is there magic to remove filters for a subdirectory or even disable mod_perl altogether? A filter can remove itself on first invocation ($f-remove). It can also remove any filter further down the filter chain ($f-next-next...-remove). So, you can write a simple filter that checks things and removes the next filter in the chain if necessary. Then you insert this filter right before the unwanted one. You can also disable mod_perl for certain request phases: http://perl.apache.org/docs/2.0/user/config/config.html#C_Perl_Handler_ or completely: http://perl.apache.org/docs/2.0/user/config/config.html#C_Enable_ Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: Linux::Smaps on RHEL
On Wed 22 Jul 2009, Jonathan Swartz wrote: I'm running into the problem using Linux::Smaps on RHEL with an Apache server on port 80. Namely, that the httpd child doesn't have permissions to look at the smaps file. Linux::Smaps: Cannot open /proc/20074/smaps: Permission denied or Linux::Smaps: Cannot open /proc/self/smaps: Permission denied I believe it is this bug: https://bugzilla.redhat.com/show_bug.cgi?id=322881 Since the redhat folks don't seem to be in a hurry to fix this, I'm wondering if anyone has discovered a workaround. (Other than, don't use RHEL - beyond my control. :)) If access is denied by the kernel there is only one solution, fix the kernel. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: rewriterule, location, and perlhandler
On Wed 22 Jul 2009, Adam Prime wrote: Eric Lease Morgan wrote: On Jul 22, 2009, at 12:05 AM, Adam Prime wrote: By first changing my Location directive to the following: Location /sandbox/semantic-web/ SetHandler perl-script PerlHandler Apache2::Alex::SemanticWeb /Location And then changing my RewriteRule to this: RewriteRule ^/etexts/id/(.*) /sandbox/semantic-web/?id=$1 [passthrough] I eliminate the need to have a file on my file system. Thank you. oss++ * mailing_lists++ If you want, it's actually possible to take this even further and remove rewrite completely. Location /etexts/id/ SetHandler perl-script PerlHandler Apache2::Alex::SemanticWeb /Location Then alter your handler's code to parse $r-uri and extract everything after $r-location. (or you can use $r-path_info, if the /etexts/id/ actually exists in your document root). That'll give you the same thing that rewrite is currently stuffing into your id argument. Never use path_info unless you are very sure it is what you want. Path_info is made for CGI scripts where $r-filename points to the file containing the script. The usage of path_info leads to action at a distance problems. Assume you have an empty directory DOCROOT/etexts/id and path_info s what you want. Later an administrator adds a file or subdirectory to that directory or removes the id directory. Suddenly path_info has changed but the handler and all libraries it uses are still the same. substr($r-uri, length $r-location) is almost always what you need. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: trying to add header field using PerlInputFilterHandler to proxy packets
On Tue 30 Jun 2009, Brandon Allgood wrote:
PerlInputFilterHandler company::AddHeader
and I wrote the following example handler
package company::AddHeader;
use strict;
use warnings;
use Apache2::Filter ();
use Apache2::RequestRec ();
use APR::Table ();
use Apache2::Const -compile = qw(OK DECLINED);
my $debug = 1;
sub handler {
my $f = shift;
# if we have already seen this do nothing
return Apache2::Const::DECLINED if $f-ctx;
# get headers
my $headers_in = $f-r-headers_in();
# add header field
$headers_in-set(Message,Hi Mom);
$f-r-headers_in($headers_in);
if($debug)
{
open FILE, /tmp/out.log or die $!;
foreach my $key (keys %{$headers_in})
{
print FILE $key = $headers_in-{$key}\n;
}
close FILE;
}
$f-ctx(1);
return Apache2::Const::OK;
}
1;
As you can see, if debugging is turned on the headers are written to
the file /tmp/out.log. The contents of out.log contains the new
header, but the requests being forwarded by the proxy don't seem to
be altered. Why is the new header not being sent?
A request level input filter is called when the request body is read.
mod_proxy does this obviously *after* sending the request header to the
backend server. Hence, adding an input header at this time is too late.
Why do you want to do that in a filter? Why don't you add it in a
request phase prior to response. Fixup would be a good place for
example.
PerlFixupHandler sub { \
use Apache2::RequestRec;\
use Apache2::Const -compile='DECLINED';\
$_[0]-headers_in-{Message}='Hi, Mom'; \
return Apache2::Const::DECLINED;\
}
Another word to your filter, did you know you can remove a filter on
first invocation? This way you can avoid the useless
return Apache2::Const::DECLINED if $f-ctx;
sub filter {
my ($f)=...@_;
# do something
$f-remove;
return Apache2::Const::DECLINED;
}
I haven't benchmarked it but I believe this is faster than return if
$f-ctx. But it only makes a difference of course if the filter is
called more than once.
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
Re: svn commit: r773881 - in /httpd/httpd/branches/2.2.x: CHANGES STATUS include/http_core.h modules/filters/mod_include.c server/config.c server/core.c
On Fri 22 May 2009, Jeff Trawick wrote: Hmmm, after trying to use what seems like a cool feature, I find that mod_perl was never taught to use the Apache 2's mod_include plug-in interface. AFAIK, that is provided by Geoff's CPAN module Apache::IncludeHook or so. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: executing PerlHandler after default_handler?
On Wed 13 May 2009, William T wrote: What I want to do is issue a redirect if the file that was requested is not on disk, and I don't want the extra stat due to an NFS mount. How about ErrorDocument 404 /modperl/handler/uri? http://httpd.apache.org/docs/2.2/mod/core.html#errordocument Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: Pass variable
On Tue 05 May 2009, Idel Fuschini wrote: I am the owner of Apache Mobile Filter project ( https://sourceforge.net/projects/mobilefilter/). After my filter detect the UserAgent of a device I want to pass the capability to other applications (cgi, php or tomcat with mod_jk) in the same web server. How can I do that ? $r-subprocess_env $r-headers_in $r-notes depends on what the other apps understand. I think subprocess_env is best choice. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: Pass variable
On Tue 05 May 2009, Idel Fuschini wrote: how can I read this variable from php and jsp application ? subprocess_env manipulates environment variables. So, if your application can read them it fetches them from that place. I did in perl code: $r-subprocess_env(XXX = $value); In cgi I see something like: REDIRECT_XXX value Somehow you managed to execute an internal redirect between the time your filter is invoked and the CGI script. Is there an ErrorDocument involved? Another way to get an internal redirect is a CGI script that prints these 2 headers: Status: 200 Location: /new/url Apache then copies all environment variables (subprocess_env) from the original request to the environment of the new (redirected) request with REDIRECT_ prepended to all names. That is why you see a REDIRECT_XXX variable. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: Undefined subroutine error
On Tue 14 Apr 2009, Roberto C. Sánchez wrote: I am currently using the latter call everywhere and it is still generating the Undefined subroutine error. Would it be possible to post your module? Perhaps you forgot the package declaration? Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: generic question on input filters
On Thu 09 Apr 2009, André Warnier wrote: Suspicion : I cannot read the content, and print nothing when I am called. Instead of accumulating the content and process it all at once at the end, should I process it chunk by chunk ? yes, as I said in my previous mail, do it chunkwise. Is it OK then when the Content-Length header is wrong (since I am changing the size of the data) ? As long as your handler does not rely on it you can ignore the Content-Length header. In fact, a request can also be sumbitted in Transfer-Encoding: chunked which means there is no CL header. The end of request body is determined by the HTTP_IN filter (ap_http_filter() in httpd/modules/http/http_filters.c) which is a protocol filter. That means it comes before your filter. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: generic question on input filters
On Thu 09 Apr 2009, André Warnier wrote: I have the impression that I am getting this Apache error : [Wed Apr 08 22:46:22 2009] [crit] [Wed Apr 08 22:46:22 2009] file http_filters.c, line 1324, assertion !bb)-list))-next == (struct apr_bucket *)((char *)(((bb)-list)) - ((long) (((char *) struct apr_bucket*)((void *)0))-link))) - ((char *) ((void *)0)) failed [Wed Apr 08 22:46:22 2009] [notice] child pid 5518 exit signal Aborted (6) Perhaps it's easier to implement a brigade based filter since this comes from here: rv = ap_get_brigade(r-input_filters, bb, AP_MODE_READBYTES, APR_BLOCK_READ, bufsiz); ... /* If this fails, it means that a filter is written incorrectly and that * it needs to learn how to properly handle APR_BLOCK_READ requests by * returning data when requested. */ AP_DEBUG_ASSERT(!APR_BRIGADE_EMPTY(bb)); Meaning your filter hasn't written anything. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: generic question on input filters
On Wed 08 Apr 2009, André Warnier wrote:
I want to write a mod_perl input filter, which essentially would
replace some characters from form-posted input on the fly, before
letting the request go to the real application.
The requests are POST, with content encoded in one long string,
URL-encoded. I am thinking of writing this as a FilterRequestHandler
filter, grabbing the content, splitting it on , URL-decoding the
so-splitted query parameters individually, then in the value
performing the character substitutions, then re-URL-encoding the
query parameters, re-joining them with , and essentially et voilà.
Basically, the type shown here :
http://perl.apache.org/docs/2.0/user/handlers/filters.html#toc_Stream
_oriented_Input_Filters
Is the above a realistic view, or am I missing something essential ?
Remember the request can consist of multiple brigades. Each filter
invocation gets one brigade. That means you have to be prepared to
store some context (store the /[^;]*$/ part in $f-ctx). Don't try to
read in the whole request. Operate on each part of it.
For example try the following PerlInputFilterHandler:
sub handler {
my ($f)=...@_;
$f-print(\n###\n);
my $buf;
while($f-read($buf, 1024)) {
$f-print(\n$buf\n);
undef $buf;
}
$f-print(\n:::\n);
return Apache2::Const::OK;
}
with this PerlResponseHandler:
sub handler {
my ($r)=...@_;
$r-content_type('text/plain');
my $buf;
while($r-read($buf, 15000)) {
$r-print($buf);
undef $buf;
}
return Apache2::Const::OK;
}
and feed it with more than 8000 bytes:
perl -e 'print xx1' | curl -qs -d @- http://localhost/iftest
You'll see multiple blocks wrapped in ... from the filter loop.
But you'll also see 2 blocks of # ... :: that mark the
filter invocations.
BTW, the buffer size in the response handler is that large because is
must be able to store a whole brigade's content due to a modperl bug,
see XXX: comment in modperl_request_read(). So, don't $f-print much
in one filter invocation.
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
Re: sending all of stderr to a process
On Sun 05 Apr 2009, E R wrote: My concern is about what happens in this scenario: 1. Set everything up (named pipe, logger process, etc.) 2. Now suppose the logger process dies (it crashes or I need to restart it) 3. Any writes through existing connections to the named pipe will just fail, and I'll lose those writes. I'm pretty sure about #3, but I might be wrong. A (named) pipe is a ring buffer within the kernel. A writer can write to it up to a certain level. Then it is full and the writer blocks or gets an EAGAIN for nonblocking writes. So there is a certain period of time when the data resides in the kernel buffer. All that data is lost when the kernel panics, loses power or similar. Then the reader reads a piece of data from the pipe and does something with it. If that process crashes between reading the data from the pipe and finishing its processing that piece is lost. So there are at least 2 periods of time when data can get lost. But is that all? It depends on your definition of sending to stderr. For example lets look at httpd. The event it wants to log happens at a certain point (in quotes because it is a short period in reality) in time. Before it goes out of stderr there is a period of time when a string is made from the event in ap_log_error(), log_error_core() or somewhere else. That perhaps requires some malloc()ing to extend a pool etc. And of course there can be a stray SIGKILL on its way to the process that arrives just in the moment before the formatted string hits stderr. That means there is a period of time when the event that is to be logged has happend but has not arrived on stderr. All I want to say here is you can't make 100% sure that every event that has to be logged is really logged. You can achieve quite a good level but never 100%. The good thing about a named pipe is that it can be connected by unrelated processes. So if the reader dies only the piece of data it is currently processing is lost. The writer can still write until the pipe is full. So, have your reader restarted fast enough. That's why I mentioned /etc/inittab. Perhaps I can't do what I want to do at the application level and I need to go to the kernel level - like write my own device driver or something, especially since I want to capture all output sent through stderr (ie. from Apache and from mod_perl and from any XS modules.) As I said before you can't achieve 100% even at kernel level. And imagine you'd want to write to a data base from kernel space. I don't say its impossible but you'd have to reimplement the whole protocol and the protocol specs may be not even published. Happy hacking! On Linux the ErrorLog of the base server is always opened on file descriptor 2 that means stderr. So, a reader that reads this stream gets everything (under normal conditions) that Apache, Perl or any XS module writes to stderr. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: sending all of stderr to a process
On Mon 06 Apr 2009, André Warnier wrote: How does it work with Apache children/threads that may be acting as a VirtualHost's having their own ErrorLog ? My experience is that anything logged by mod_perl modules running in a VirtualHost using explicit warn() or $r-log_error() or $r-log-info() etc.. ends up in the VirtualHost's ErrorLog, but some run-time messages like Use of uninitialized value in concatenation (.) or string at .. end up in the main Apache error log. How does that come about ? $r or $s based logging goes to the VHosts ErrorLog, yes. But the original poster worried about all of STDERR come it from perl, xs or whereever else. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: sending all of stderr to a process
On Sat 04 Apr 2009, E R wrote: One operational issue with this is that if the logger process dies, existing httpd processes will get a SIGPIPE signal on subsequent writes. Restarting the logger process does not restore the lost connections. On Fri, Apr 3, 2009 at 12:38 PM, Torsten Foertsch torsten.foert...@gmx.net wrote: On Fri 03 Apr 2009, E R wrote: I want to send all error output to a process for specialized logging. I know that you can use CustomLog with a pipe. However, will that capture output from things like 'print STDERR ...' or if an XS module directly writes to stderr? I need that output to also go to my logging process. Does anyone know if there an Apache module which does this? Give a named pipe's name to ErrorLog outside any virtual host. Then have your process read from that pipe. At least with prefork and worker MPMs apache ignores SIGPIPE: httpd-2.2.11/server/mpm/prefork/prefork.c:406: (similar in worker.c) #ifdef SIGPIPE sa.sa_handler = SIG_IGN; if (sigaction(SIGPIPE, sa, NULL) 0) ap_log_error(APLOG_MARK, APLOG_WARNING, errno, ap_server_conf, sigaction(SIGPIPE)); #endif And later on it does not check the return code when writing the log: httpd-2.2.11/server/log.c:682: apr_file_puts(errstr, logf); apr_file_flush(logf); Back in 1999-2000 apache had a reliability problem with piped loggers. Ever since I log to named pipes. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: sending all of stderr to a process
On Fri 03 Apr 2009, E R wrote: I want to send all error output to a process for specialized logging. I know that you can use CustomLog with a pipe. However, will that capture output from things like 'print STDERR ...' or if an XS module directly writes to stderr? I need that output to also go to my logging process. Does anyone know if there an Apache module which does this? Give a named pipe's name to ErrorLog outside any virtual host. Then have your process read from that pipe. At shell level: mkfifo /path/to/PIPE your_program /path/to/PIPE in httpd.conf: ErrorLog /path/to/PIPE I'd very much recommend to monitor your_program and have it restarted automatically if it crashes. The simplest way to do that is to start it via /etc/inittab. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: Help needed for mod_perl 1.31-rc6 and Apache::Bootstrap
On Sun 29 Mar 2009, Fred Moyer wrote: Need one more for a release $ make test cp lib/Apache/Bootstrap.pm blib/lib/Apache/Bootstrap.pm PERL_DL_NONLAZY=1 /usr/bin/perl -MExtUtils::Command::MM -e test_harness(0, 'blib/lib', 'blib/arch') t/*.t t/00-load.t ... 1/11 # Testing Apache::Bootstrap 0.05-rc2, Perl 5.01, /usr/bin/perl Apache::Test version is 1.31, minimum version required is 1.32, tests will be skipped t/00-load.t ... ok t/pod-coverage.t .. ok t/pod.t ... ok All tests successful. Files=3, Tests=13, 1 wallclock secs ( 0.04 usr 0.01 sys + 0.73 cusr 0.06 csys = 0.84 CPU) Result: PASS This is Suse Linux 11.1. Kernel: Linux opi 2.6.27.19-3.2-default #1 SMP 2009-02-25 15:40:44 +0100 x86_64 x86_64 x86_64 GNU/Linux mod_perl: the current threading branch with all patches for trunk merged. The one thing that puzzles me is Apache::Test 1.32. My current copy of trunk shows: $ svn info clean-trunk/Apache-Test/lib/Apache/Test.pm Path: clean-trunk/Apache-Test/lib/Apache/Test.pm Name: Test.pm URL: https://svn.apache.org/repos/asf/perl/Apache-Test/trunk/lib/Apache/Test.pm Repository Root: https://svn.apache.org/repos/asf Repository UUID: 13f79535-47bb-0310-9956-ffa450edef68 Revision: 759858 Node Kind: file Schedule: normal Last Changed Author: gozer Last Changed Rev: 607695 Last Changed Date: 2007-12-31 09:00:09 +0100 (Mon, 31 Dec 2007) Text Last Updated: 2008-09-30 09:52:57 +0200 (Tue, 30 Sep 2008) Checksum: 61e9675a4f820b7792c2f8313dd3d97c And that file contains: $VERSION = '1.31'; What did I miss? Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: who's putting that pre tag in the output...?
On Mon 23 Mar 2009, Iosif Fettich wrote:
So it seems to be very, very easy. Still, when using the above
receipt like
RewriteEngine on
RewriteCond %{REQUEST_URI} !-U
RewriteRule ^\/(.+) http://OLDDOMAIN.COM/$1 [QSA,P]
The engine tries to resolve the request uri with a subrequest. Hence it
goes through this rule twice. Try to add NS (no subrequest) to the
flags [QSA,P,NS].
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
Re: who's putting that pre tag in the output...?
On Mon 23 Mar 2009, Iosif Fettich wrote:
Or is the [P]roxy flag not working as it should or as I expect it to
? It seems to work fine for the subrequest (status=200) ...?
This is exactly the problem. The 404 is normally generated in the
response phase from the default response handler. The subreq lookup
won't check that. It does not run the subreq but only checks if after
fixup there is still no error.
So the simplest solution for you would be an ErrorDocument I think.
Something like this:
ErrorDocument 404 /-/404 - put an url here you are not using otherwise
RewriteRule ^/-/404 http://...%{REDIRECT_URI} [P]
Write a simple printenv and activate it as ErrorDocument first to get
the right environment variable names (not sure if it is REDIRECT_URI).
If you use mod_include that would be
!--#printenv --
in an .shtml file.
This solution should work for GET/HEAD. POST requests are still a
problem. If you use them we'll make it work without ErrorDocument. BTW,
why don't you use the file lookup (-f). That is much easier. Is your
uri-file mapping so complicated?
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
to the mailing list admin
Hi, each of my last mailings to the list was answered by the following failure notice: - The following message to sbirl+excha...@concept.temple.edu was undeliverable. The reason for the problem: 5.1.0 - Unknown address error 550-'5.7.1 torsten.foert...@gmx.net... Access denied' - Please remove this recipient from the list. Thanks, Torsten
Re: who's putting that pre tag in the output...?
On Mon 23 Mar 2009, Iosif Fettich wrote: Using the (obvious...!) REDIRECT_URL as you suggested works! :) An ErrorDocument is an internal redirect. These REDIRECT_... environment variables are copied from the previous ($r-prev) request's $r-subprocess_env just by copying everything and prepending REDIRECT_ to each key. So if the original request has an environment variable named REQUEST_URI the error document should have a REDIRECT_REQUEST_URI, see rename_original_env() in httpd-x.y/modules/http/http_request.c. Since REQUEST_URI is the standard CGI environment variable (see ap_add_cgi_vars() httpd-x.y/server/util_script.c) I'd take REDIRECT_REQUEST_URI. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: generated report (was Re: mp2 hangs when running make test)
On Sat 21 Mar 2009, Oliver Block wrote: *** mod_perl version 2.04 And how about the current development version? I use a self compiled apache 2.2.11 plus http://svn.apache.org/repos/asf/perl/modperl/branches/threading/ with Suse 11.1. I think you are at 11.0? If you like I can send you the SRPMs or at least the spec files. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: mp2 hangs when running make test
On Fri 20 Mar 2009, Oliver Block wrote: [warning] the client side drops 'root' permissions and becomes 'nobody' not sure, but I'd retry the complete build/test (beginning from extracting the TAR) without superuser privileges. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: Capturing STDERR for use in log handler
On Tue 17 Mar 2009, Tom Spencer wrote:
So far so good, it adds the information I've got to the database
quite nicely. But the one more thing I really need is all the STDERR
that was generated while processing the request. Recipe 6.10
provides a method for outputting STDERR to a different file but I'd
like to grab it and put it into my database.
First thing, you have to specify your understanding of all the STDERR
in more detail. Is it
1) the warn and die output
2) all that perl writes to STDERR
3) something that XS modules or rather libraries used by XS modules
write to filedescriptor 2
4) the stuff that apache logs to the virtual hosts errorlog
Here are some ideas how to solve these problems:
1) set $SIG{__WARN__} and $SIG{__DIE__} or implement CORE::GLOBAL::warn
and CORE::GLOBAL::die
2) reopen STDERR to a memory filehandle, eg:
perl -Mstrict -e '
# list our open fds to see what is going on
system ls -l /proc/$$/fd;
# save fd2 as SAVEERR
open SAVEERR, STDERR or die Cannot dup: $!;
close STDERR;
# make sure that nobody else can open fd2 by chance
open my $acquirefd2, , /dev/null or die acquire: $!;
# check it
system ls -l /proc/$$/fd;
my $err;
# open memory filehandle
open STDERR, , \$err or die open: $!;
# this output must not differ from the previous
system ls -l /proc/$$/fd;
# some printout to STDERR and a warning
# nothing appears on the terminal
print STDERR print to STDERR\n;
warn warning;
# print the error output to STDOUT surrounded by
print $err\n;
close STDERR;
system ls -l /proc/$$/fd;
# release fd2
close $acquirefd2;
# redirect STDERR to the original fd2
open STDERR, SAVEERR or die restore: $!;
close SAVEERR;
# check it
system ls -l /proc/$$/fd;
# these will appear on the terminal
print STDERR print to STDERR\n;
warn warning;
'
This is what I see (a bit commented):
total 0
lrwx-- 1 r2 users 64 2009-03-18 12:32 0 - /dev/pts/9
lrwx-- 1 r2 users 64 2009-03-18 12:32 1 - /dev/pts/9
lrwx-- 1 r2 users 64 2009-03-18 12:32 12 - socket:[12916]
lrwx-- 1 r2 users 64 2009-03-18 12:32 13 - socket:[12917]
# fd2 points to my terminal
lrwx-- 1 r2 users 64 2009-03-18 12:32 2 - /dev/pts/9
lr-x-- 1 r2 users 64 2009-03-18 12:32 3 - pipe:[10303]
l-wx-- 1 r2 users 64 2009-03-18 12:32 4 - pipe:[10303]
lr-x-- 1 r2 users 64 2009-03-18 12:32 5 - pipe:[31529]
lrwx-- 1 r2 users 64 2009-03-18 12:32 7 - socket:[12899]
total 0
lrwx-- 1 r2 users 64 2009-03-18 12:32 0 - /dev/pts/9
lrwx-- 1 r2 users 64 2009-03-18 12:32 1 - /dev/pts/9
lrwx-- 1 r2 users 64 2009-03-18 12:32 12 - socket:[12916]
lrwx-- 1 r2 users 64 2009-03-18 12:32 13 - socket:[12917]
# now it points to /dev/null
lrwx-- 1 r2 users 64 2009-03-18 12:32 2 - /dev/null
lr-x-- 1 r2 users 64 2009-03-18 12:32 3 - pipe:[10303]
l-wx-- 1 r2 users 64 2009-03-18 12:32 4 - pipe:[10303]
# but fd5 (SAVEERR) points to my terminal
lr-x-- 1 r2 users 64 2009-03-18 12:32 5 - /dev/pts/9
lr-x-- 1 r2 users 64 2009-03-18 12:32 6 - pipe:[31546]
lrwx-- 1 r2 users 64 2009-03-18 12:32 7 - socket:[12899]
total 0
lrwx-- 1 r2 users 64 2009-03-18 12:32 0 - /dev/pts/9
lrwx-- 1 r2 users 64 2009-03-18 12:32 1 - /dev/pts/9
lrwx-- 1 r2 users 64 2009-03-18 12:32 12 - socket:[12916]
lrwx-- 1 r2 users 64 2009-03-18 12:32 13 - socket:[12917]
lrwx-- 1 r2 users 64 2009-03-18 12:32 2 - /dev/null
lr-x-- 1 r2 users 64 2009-03-18 12:32 3 - pipe:[10303]
l-wx-- 1 r2 users 64 2009-03-18 12:32 4 - pipe:[10303]
lr-x-- 1 r2 users 64 2009-03-18 12:32 5 - /dev/pts/9
lr-x-- 1 r2 users 64 2009-03-18 12:32 6 - pipe:[31554]
lrwx-- 1 r2 users 64 2009-03-18 12:32 7 - socket:[12899]
# the framed output
print to STDERR
warning at -e line 19.
total 0
lrwx-- 1 r2 users 64 2009-03-18 12:32 0 - /dev/pts/9
lrwx-- 1 r2 users 64 2009-03-18 12:32 1 - /dev/pts/9
lrwx-- 1 r2 users 64 2009-03-18 12:32 12 - socket:[12916]
lrwx-- 1 r2 users 64 2009-03-18 12:32 13 - socket:[12917]
# still the same: fd2 points to /dev/null, fd5 to my terminal
lrwx-- 1 r2 users 64 2009-03-18 12:32 2 - /dev/null
lr-x-- 1 r2 users 64 2009-03-18 12:32 3 - pipe:[10303]
l-wx-- 1 r2 users 64 2009-03-18 12:32 4 - pipe:[10303]
lr-x-- 1 r2 users 64 2009-03-18 12:32 5 - /dev/pts/9
lr-x-- 1 r2 users 64 2009-03-18 12:32 6 - pipe:[31561]
lrwx-- 1 r2 users 64 2009-03-18 12:32 7 - socket:[12899]
total 0
lrwx-- 1 r2 users 64 2009-03-18 12:32 0 - /dev/pts/9
lrwx-- 1 r2 users 64 2009-03-18 12:32 1 - /dev/pts/9
lrwx-- 1 r2 users 64 2009-03-18 12:32 12 - socket:[12916]
lrwx-- 1 r2 users 64 2009-03-18 12:32 13 - socket:[12917]
# here it is reverted, fd2 is the terminal, fd5 is something else
# (due to the ls command)
lrwx-- 1 r2 users 64 2009-03-18 12:32 2 - /dev/pts/9
lr-x-- 1 r2 users 64 2009-03-18 12:32 3 - pipe:[10303]
l-wx-- 1 r2 users 64 2009-03-18 12:32 4 - pipe:[10303]
lr-x-- 1 r2 users 64 2009-03-18 12:32 5
BerkeleyDB solution
Hi,
this is not a problem but a solution. I know some of you use the
BerkeleyDB module to store data. Recently I have tried to use UTF8 keys
and failed. When reading back keys I sometimes got character strings
sometimes octet strings. I had used the following 2 filters to ensure
the data in the database is octet strings and the data I get back are
character strings:
$db-filter_fetch_key(sub { $_=Encode::decode('utf8', $_) });
$db-filter_store_key(sub { $_=Encode::encode('utf8', $_) });
The problem is BerkeleyDB doesn't reset the UTF8 bit when storing data
to @_ variables as in c_get() or db_get(). One possible solution is
$db-filter_fetch_key(sub {
Encode::_utf8_off($_);
$_=Encode::decode('utf8', $_);
});
The other/better one is to fix it in BerkeleyDB.xs. This is what the
attached patch does. I have sent it to the author, Paul Marquess. Here
is his reply:
On Tue 10 Mar 2009, Paul Marquess wrote:
Your patch looks fine and should be ok to include in my development
copy without any changes.
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
--- BerkeleyDB.xs~ 2009-02-18 21:31:46.0 +0100
+++ BerkeleyDB.xs 2009-03-06 14:38:04.0 +0100
@@ -430,7 +430,10 @@
#define getInnerObject(x) ((SV*)SvRV(sv))
#endif
-#define my_sv_setpvn(sv, d, s) (s ? sv_setpvn(sv, d, s) : sv_setpv(sv, ) )
+#define my_sv_setpvn(sv, d, s) do { \
+ s ? sv_setpvn(sv, d, s) : sv_setpv(sv, ); \
+ SvUTF8_off(sv); \
+ } while(0)
#define GetValue_iv(h,k) (((sv = readHash(h, k)) sv != PL_sv_undef) \
? SvIV(sv) : 0)
Re: PerlConfigRequire and die
On Thu 26 Feb 2009, Clinton Gormley wrote:
Why don't you do that in a Perl block?
How is this different from doing it in the startup.pl file?
Yes, I see, the 2nd run of the Perl code doesn't have STDERR either.
So, you can open a startup logfile in perl and override
CORE::GLOBAL::warn/die or install a $SIG{__WARN__/__DIE__} handler. Or
even open that logfile at start of your startup.pl on file descriptor 2
and close it at the end.
Perl
use Apache2::ServerUtil;
if( Apache2::ServerUtil::restart_count1 ) {
open my $olderr, '', \*STDERR;
close STDERR;
open STDERR, '', '/tmp/myerr';
# initialize your app here
warn huhu;
print STDERR haha;
# end of init
close STDERR;
open STDERR, '', $olderr;
}
/Perl
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
Re: PerlConfigRequire and die
On Thu 26 Feb 2009, Clinton Gormley wrote:
Is there any way I can:
- cause the error to be reported properly
- force apache not to start
eval {load_application(); 1} || do { warn $@; die };
Unfortunately, no. That still just dies with the obscure error
message. It seems that STDERR only gets flushed in a later stage of
the startup process.
What I was thinking was, could there be some way of saying:
eval { load_application();1}
|| do {
warn $@;
force_apache_to_quit_startup_once_stderr_flushed();
}
Why don't you do that in a Perl block?
Be aware that this code is executed twice at startup. Check
Apache2::ServerUtil::restart_count to prevent that.
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
Re: What can a child_init do?
On Thu 19 Feb 2009, cr...@animalhead.com wrote:
Is there anything a mod_perl2 child_init phase can do to call
attention to a problem?
I moved a block of code from a post_config handler to a new
child_init handler, without thinking much about the niceties of the
move. The code contained a couple of 'die' statements, which I trust
would prevent an Apache startup if executed in a post config handler.
Q1: WOULD 'DIE' IN A POST_CONFIG HANDLER ABORT AN
APACHE2 STARTUP?
yes. Place the following line in your httpd.conf:
PerlPostConfigHandler sub {die q{huhu}}
and you'll see something like this in your error log:
[Thu Feb 19 08:51:21 2009] [error] huhu at (eval
50) line 1.\n
Configuration Failed
and of course no running httpd.
In the child_init handler, an executed 'die' did nothing noticeable.
Specifically the message did not end up in the error_log. In the
mod_perl2 doc pages, child_init is described as being 'of type void',
which another page says means that the return value doesn't matter.
I will change the 'die' to a '$s-log_error', and return the
nastiest- sounding Apache return code I can find, in hopes that some
future Apache might notice it.
Q2: IS THERE ANYTHING A CHILD_INIT PHASE CAN DO TO
ABORT ITS CHILD PROCESS, AND THUS CALL ATTENTION
TO A SERIOUS PROBLEM?
PerlChildInitHandler sub {warn q{about to commit suicide}; kill 9, $$}
or
PerlChildInitHandler sub {warn q{exiting now}; CORE::exit -1}
But you should not need that. Because when apache needs another worker
it will start one (or more). If they exit immediately apache will keep
starting new ones.
Also, such things are bad for your users experience. The server accepts
the connection but doesn't respond. The request times out without
receiving a single byte. I think it's better to serve at least one
request and inform the user about temporary problems. You can ask the
current worker to die when the current request is done by
$r-child_terminate (or was it terminate_child?).
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
Re: Alias and PerlOutputFilterHandler in virtual host configuration
On Sun 15 Feb 2009, Roger Munk wrote: I have the following directives in my sites-available/default configuration: DocumentRoot /var/www/ Alias /portal/ /var/www/drupal6/ PerlRequire /etc/apache2/ModPerl/TE/ST.pm Location /portal/ SetHandler modperl PerlOutputFilterHandler TE::ST /Location When the SetHandler and PerlOutputFilterHandler are commented out, the redirection from /portal to /var/www/drupal6 works correctly. However, when the SetHandler and PerlOutputFilterHander are enabled I get the following error messages: [Mon Feb 16 22:43:40 2009] [error] [client 192.168.1.2] Attempt to serve directory: /var/www/drupal6/, referer: http://10.0.0.1/portal/ [Mon Feb 16 22:43:40 2009] [error] [client 192.168.1.2] script '/var/www/index.php' not found or unable to stat, referer: http://10.0.0.1/portal/ Why would the SetHandler and PerlOutputFilterHandler stop the Alias directive from working correctly? It does not. You see the first message? That is the result of the alias. The thing that confuses me is what do you want to achieve by SetHandler modperl but leaving out the PerlResponseHandler? If your page is generated by php and you only want to postprocess it on its way out then drop the SetHandler. The PerlOutputFilterHandler does not need a SetHandler modperl. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: Alias and PerlOutputFilterHandler in virtual host configuration
On Sun 15 Feb 2009, Roger Munk wrote:
On Sun, Feb 15, 2009 at 5:12 PM, André Warnier a...@ice-sa.com wrote:
see http://perl.apache.org/docs/2.0/api/APR/Table.html#C_get_
(you only get the first one this way)
my @cookies = $f-r-headers_out-get(Set-Cookie);
would get you an array with both.
Thanks, when I tried receiving the cookies in an array, using the
code below, I only see the second cookie now in the output. Since
both cookies are set with seperate Set-Cookie headers are they
getting overwritten at some point, or is there a problem with my
code?
unless ($f-ctx)
{
my @Cookies = $f-r-headers_out-get(Set-Cookie);
foreach my $Cookie (@Cookies)
{
$Cookie =~ s/expires.*GMT\;//g;
$f-r-headers_out-set(Set-Cookie = $Cookie);
}
$f-ctx(1);
}
I am about to say RTFM but being polite: Please read the documentation
mentioned above more thouroughly. APR::Table provides an add method.
And before you ask your next question How can I get rid of the
original cookies? read that piece again, ;-)
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
Re: Checking a webserver
On Wed 28 Jan 2009, André Warnier wrote: I am looking for a debugging tool that would be able to repeatedly issue HTTP requests to one or more URLs, and log any errors. Preferably in Perl, and it must run on a Windows workstation. Don't know about windows and its not perl but I use ab (comes with httpd) and flood: http://httpd.apache.org/test/flood/ Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: PerlSetVar behaviour within Limit directive
On Fri 23 Jan 2009, Pat Downey wrote: Limit GET PerlSetVar CrowdAllowedGroups reader,writer Require valid-user /Limit LimitExcept GET PerlSetVar CrowdAllowedGroups writer Require valid-user /LimitExcept /LocationMatch What I'd like is for members of the reader or writer groups to have GET access and only members of the writer group to have everything else (mainly for POST) access. The problem seems to be that the 'PerlSetVar CrowdAllowedGroups writer' is overwriting the 'PerlSetVar CrowdAllowedGroups reader,writer' call for all requests, not just the method that I'm trying to limit each setting to. Then, how about PerlSetVar CrowdAllowedGroups reader,writer Require valid-user LimitExcept GET PerlSetVar CrowdAllowedGroups writer /LimitExcept Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: Initializing Sleepycat::DbXml (Berkeley, Oracle) objects in startup.pl
On Sun 18 Jan 2009, Michael Ludwig wrote:
From perldoc perlthrtut: In this model each thread runs in its own
Perl interpreter, and any data sharing between threads must be
explicit. This does not sound to me as if there is a significant
advantage over spawning child processes, at least not on UNIX.
Hmm. Not sure what to make of this threaded Perl.
In fact, it is worse than fork()ing off unix processes because the
interpreter data is completely copied while a new interpreter is
created. On the other hand a forked process copies only those pages
that are written to.
I haven't used Sleepycat::DbXml but I use BerkeleyDB in production with
prefork, see Apache2::Translation on CPAN. What I have learned is, you
can't open the BDB env in the parent apache. When I wrote that piece I
also had problems when I opened the env in the parent and then closed
it before forking children. Perhaps they are gone in a newer version.
I think with prefork opening the env in a ChildInit handler should work.
I ended up with the connect-at-first-usage approach
unless( $connected ) {
connect();
$connected=1;
}
because I test almost all my modules also with a threaded MPM and the
ChildInit hook is run once per process not once per Perl interpreter.
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
Re: Initializing Sleepycat::DbXml (Berkeley, Oracle) objects in startup.pl
On Mon 19 Jan 2009, Michael Ludwig wrote: In fact, it is worse than fork()ing off unix processes because the interpreter data is completely copied while a new interpreter is created. On the other hand a forked process copies only those pages that are written to. Thanks. If I'm understanding correctly and this is true (and not considering SEGVs), the total memory cost for a mod_perl on a worker MPM should outweigh that for mod_perl on a prefork MPM, which seems to not be aligned with what Craig has observed. Of course it depends on your setup. If you configure only a small number of interpreters then the overall memory footprint can be lower then with prefork where each process runs its own interpreter. But the effect of the copied interpreters outweighs this very soon. Measuring memory consumption on Linux is not simple. You can't simply add up all RSS or VSZ figures or something like this. I know of 2 approaches that work more or less well. 1) call mlockall in the parent and then use /proc/PID/smaps (or Linux::Smaps) to measure shared memory 2) take a system with enough RAM and nothing else to do, start apache with different numbers of children and watch the free memory with vmstat while doing so. Then you can estimate the amount of memory that is consumed by one apache process. See the attached picture for example. In order to be on the safe side as far as Berkeley DB is concerned, you'd have to make sure up front that the environment is safe to use, which is done by running recovery (DB_RECOVER flag, DB_RUNRECOVERY error code) in a single thread or process that has exclusive access to the environment. You can run recovery in a separate process spawned at startup by the parent. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net attachment: experiment-scrambled.png
Re: Multiple mod_perl 2.0 installations on one Linux machine
On Mon 19 Jan 2009, Adam Prime wrote: I thought you could build more than one and load the one you want at runtime. Isn't that what Red Hat does? If redhat's shipping more than one mpm, they've patched httpd. Probably not httpd. At least Suse has not. Instead they configure and build a prefork httpd and a worker httpd one by one and then merge the result. The apache API does not depend on the chosen MPM. So, almost certainly you can build modules (including mod_perl) with one MPM and use it with another. Usually I build mod_perl against a httpd with worker MPM and TEST is with both worker and prefork. So far I have never found something that went bad only because of the other MPM. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: Separate Error log for each virtual host
On Fri 16 Jan 2009, fREW Schmidt wrote:
I would like to configure apache such that the errors for a specific
virtual host get logged in their own file. I tried something like
this:
VirtualHost *:8080
ErrorLog C:/location/of/acd/logs/error.log
Perl
use lib 'C:/location/of/acd';
$ENV{MODE} = 'development';
/Perl
ScriptAlias / C:/locattion/of/acd
Location /
SetHandler perl-script
PerlHandler ACD::Dispatch
Order allow,deny
Allow from all
/Location
/VirtualHost
But it seems to be ignoring my ErrorLog directive and still puts the
logs in the regular place. What am I doing wrong?
Are you wondering why all your warn() output is directed to only one log
file? If yes, then try the logging functions in Apache2::Log instead.
I don't know about windows but on a UNIX-like system the default
ErrorLog is opened on file descriptor 2 and that happens to be STDERR.
That's why all the perlish warn()ings go to that file.
You will notice that the methods in Apache2::Log all require a request
or server object. This is necessary to decide into which error log the
message should be written. Since the perl warn() lacks that information
it cannot log to a specific file.
You can override the warn() by either supplying a $SIG{__WARN__} handler
or by implementing *CORE::GLOBAL::warn=sub {...}. That function can
then check if Apache2::RequestUtil-request or
Apache2::ServerUtil-server returns something useful and the use that
object with Apache2::Log.
If your Apache error messages all go to the same file then your
configuration is wrong. If you use named virtual hosts you have perhaps
forgotten the ServerName directive (as Adam pointed out) or your
requests lacks the Host-header. I think Apache complains in that case
at startup. Turn on LogLevel debug. For IP/port based vhost you don't
need either.
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
Re: generic enquiry about rendering zipped content
On Fri 16 Jan 2009, André Warnier wrote:
What exact form would a HTTP response need to have, for the browser
to correctly interpret that the response it is getting is a document
(for example an OpenOffice document or an email in eml format), but
which has been zipped for transmission ?
What I would like to happen is that the browser receives the zipped
document, but instead of proposing to open it as a zip file or save
it to disk as a zip file, unzips it and handles it properly as per
the content of the zipped response.
I know that this has to do with the Content-Type and
Content-disposition and Content-encoding (Content-type-encoding ?)
headers, but do not know exactly what are the caveats or rules or
things that work with one browser and not the other kind of stuff.
Does anyone know ?
If more context is required :
A server-side application stores on the one hand original documents
of different kinds, zipped, and on the other hand meta-data about
these documents in a database and search-engine.
The user can search the database, and obtain a response html page, on
which appear icons representing the original documents. When
clicking on such an icon, the user should receive the original
document. It seems a pity, considering that the document is already
stored zipped at the server side, to unzip it before sending it to
the browser, so that the browser would know that it is not a
zip-file, but an OpenOffice document.
The plan is thus to have, underneath the icons, a link which invokes
an application on the server (Apache module, mod_perl module,
cgi-bin,..), which retrieves the zipped original, composes the
appropriate HTTP headers (wich it can do only based on info stored in
the db), and sends the document in the appropriate way to the
browser, in the zipped format. The filename of the zipped document on
the server gives no clue as to its content. Only the meta-data in the
db indicates that.
Thanks in advance for any insights, advice, pointers etc..
You can send the file in DEFLATE format (rfc1951) or in GZIP format
(1952). The latter happens to be the output of the gzip program. Winzip
Co use a different format! I think I do something like you want to.
use APR::Finfo ();
use APR::Const -compile=qw/FINFO_NORM/;
if( $r-finfo-filetype and
$r-headers_in-{'Accept-Encoding'}=~/gzip/ and
-f $r-filename.'.gz' ) {
$r-filename($r-filename.'.gz');
$r-content_encoding('gzip');
$r-finfo(APR::Finfo::stat($r-filename, APR::Const::FINFO_NORM,
$r-pool));
$r-headers_out-add(Vary='Accept-Encoding');
}
You have to check if the browser accepts gzip format. Set the
Content-Encoding header and the Vary header (for proxies).
mod_negotiation (MultiViews) provides a similar functionality.
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
Re: [mp2] a way to map a uri to a physical path
On Thu 15 Jan 2009, Foo JH wrote:
Is there a generic method so that given any uri as a parameter, the
library can do the math and return the physical path?
If you look for the filename for $r-uri, that means the uri of the
current request then $r-filename holds that after the map-to-storage
phase.
If you need a general method to map an arbitrary URI to a filename then
it is a subrequest:
my $subr=$r-lookup_uri($uri);
if( $subr-status == Apache2::Const::OK and
-f $subr-filename ) {
$filename=$subr-filename;
}
But keep in mind that it may result in quite unusual filenames if you
use the mod_proxy handler for example. Also, the file may not exist.
The status 404 is generated only in the response phase. Further, you
must perhaps check $subr-path_info for being empty.
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
Re: Flushing output and preserving headers
On Sat 03 Jan 2009, Kate Yoak wrote:
$r-headers_out-set('Content-Type' = image/gif);
use $r-content_type(...)
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
Re: HTTP Response Headers fixup
On Thu 01 Jan 2009, André Warnier wrote: It works perfectly, even without having the filter remove itself (which I did not know how to do). http://perl.apache.org/docs/2.0/api/Apache2/Filter.html#C_remove_ Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: HTTP Response Headers fixup
On Thu 01 Jan 2009, André Warnier wrote: Unfortunately the Content-Type header is a different beast. Inside Apache it is not only a response header, but a more complex data type. You can set a different Content-Type header with mod_headers, but since the internal structure remains unchanged it will be overwritten again by Apache. As a result I see no way to change an existing character set in a Content-Type header. Try to create a request output filter that sets $r-content_type, removes itself and returns DECLINED. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: Keep-Alive, Apache processes and $r-connection
On Sun 21 Dec 2008, André Warnier wrote: Am I right, or horribly wrong, to assume that in case of a keep-alive client connection, the multiple requests sent over that connection will be processed by the same Apache process (child or thread) ? I am assuming the above, from the fact that there exists a $request-connection-notes table. It seems that it would make little sense for such a structure to exist otherwise. Yes, and there is also a $r-connection-pnotes hash to store perl data structures. Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: sub-requests, lookup_uri and that kind of stuff
On Mon 22 Dec 2008, André Warnier wrote:
I could do this the hard way by using LWP to just issue a new HTTP
request to localhost and get the content, before I go on with my
current request.
But I figure there must be an easier and more efficient way, no ?
Looking at the explanations for sub-requests, lookup_uri etc.., I get
the impression however that calling these, pretty much terminates the
current request, and replaces it by the one I'm calling.
Which is not what I want.
I want to continue processing the current request but using the
response from my other request.
Look at what mod_include does to get an impression about subrequests.
You can use subrequests in 3 ways:
1) lookup_uri or similar creates the subreq and runs it up to fixup.
Then you can use any information the subreq has gathered so far, that
means authentication, finfo etc.
2) run it to include the output of the subreq into the current document
more or less unmodified. To do that you create the subreq, install the
necessary output filters and run() it.
3) run it to process the output in some other way. That can be done for
example this way:
my $content='';
my $subr=$r-lookup_uri( $uri );
$subr-add_output_filter( sub {
my ($f, $bb) = @_;
while (my $e = $bb-first) {
$e-read(my $buf);
$content.=$buf;
$e-delete;
}
return Apache2::Const::OK;
});
$subr-run;
my $rc=$subr-status;
my $ct=$subr-content_type;
Now, $content holds the response body, $rc the HTTP status code and $ct
the subreq's content type.
In all cases a subreq returns control to the calling request. An
internal_redirect on the other hand almost completely takes over the
request without return to the calling req.
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
Re: content_type in SSI- or directly-called mp2 scripts
On Wed 17 Dec 2008, cr...@animalhead.com wrote:
My mod_perl2 scripts mostly fall into 2 categories:
1) scripts that are called by URL location, and generate complete
content-pages
2) scripts that are called by SSI include virtual sequences
in .html files, and generate part of a page
In some cases scripts of type 1 directly call scripts of type 2 to
generate parts of their pages.
(Call this case 3).
My questions are about when to call $r-content_type('text/html')
Such a call is a good idea in category 1, right?
Each request has a content-type assigned no matter if you set it in your
script or not. Normally (for static files), that header if figured out
by one of the mime-modules (mod_mime, mod_mime_magic) in the
type-checker-phase of the request cycle. But usually it's a good idea
to set/reset it in your content handler prior to sending any data to
the client because that handler knows better what it is going to send.
Such a call probably should not be made by the directly-called script
in case 3, right?
Apache probably can't even tell that a new script has gotten into the
act, the calling
and called script are both in the undifferentiated sea of mod-perl
code, right?
In an SSI-invoked script (category 2) is a content_type call a)
required, b) good practice,
or c) a bad idea?
In all cases it is good practice. Virtual include uses subrequests. That
means when the INCLUDE filter sees a vinclude it stops the output of
the current request, creates a subrequest for the wanted document, runs
that document sending the output to the client but ignoring all headers
save the status. When the subrequest is done the output of the main
request is resumed. So, in your cases 2 and 3 the Content-Type header
is ignored. The client sees only the content-type of the main request.
If your included script wants to know the content-type of the main
request it can use $r-main-content_type (perhaps to adopt the
character set).
Torsten
--
Need professional mod_perl support?
Just hire me: torsten.foert...@gmx.net
Re: Problem - Missing input data
On Wed 17 Dec 2008, mod_perl User wrote: I am facing problem in mod_perl2.0 while opening an handler Code as follows, my $r = Apache2::Request-new(shift, POST_MAX = 10 * 1024 * 1024); my $status = $r-parse(); Getting $status as 'Missing Input Data' I am not an expert in libapreq but I have seen this when there is no request body (GET or empty POST). Use instead $req=Apache2::Request-new($r); $r-discard_request_body; # read in the complete body $status=$r-method eq 'POST' ? $req-body_status || $req-args_status : $req-args_status; Torsten -- Need professional mod_perl support? Just hire me: torsten.foert...@gmx.net
Re: Question about Apache2::ServerUtil and PerlCleanupHandler
On Fri 05 Dec 2008, Kostas Chatzikokolakis wrote: I'd like to ask what is the intended behaviour of Apache2::ServerUtil-server-push_handlers(PerlCleanupHandler = ...) compared to Apache2::RequestUtil-request-push_handlers(PerlCleanupHandler = ...) On my Ubuntu 8.10 (mod_perl 2.0.4, apache 2.2.9) cleanup handlers installed on $s seem to be called after *every* request but handlers installed on $r are called only *once* after the *current* request. On a RedHat ES 5 (mod_perl 2.0.2, apache 2.2.3) cleanup handlers on $s look like they are never called. I found and old list message suggesting that cleanup handlers on $s are called when the server is terminated, but this doesn't seem to be the case (and looks strange since we have PerlChildExitHandler for that). In the documentation of handlers, there is no mention of where each handler must be installed and whether it makes a difference. NOTE: on my Ubuntu, a cleanup handler installed on $s is called only if there are also other types of handlers on $s. In the Changelog of the *threading branch* I found the following message Now correctly invokes PerlCleanupHandlers, even if they are the only handler type configured for that request [Torsten Foertsch] this comes from r594609 but it was not merged for 2.0.4. This maybe happens also for $r handlers, but on my setup $r has always at least a PerlResponseHandler so I'm not sure. Apache request configuration is performed in 2 steps. The first occurs at startup the second at runtime. At startup one configuration object per virtual server is created. At runtime a request inherits from that config but other request specific configuration sources (Location blocks, .htaccess files etc) are merged. $s-push_handlers as well as $s-add_config modify the server configuration object while $r-... modify the runtime request configuration. The server methods are intended to be used only at startup time (up to PerlPostConfig, perhaps PerlChildInit works too). Later the server configuration must be read-only or you risk segfaults at least on a threaded MPM. The PerlCleanup phase is an artificial thing that is implemented as a pool cleanup function on the request pool. But it needs to be registered on the request pool to be run. If there is no Perl handler in the request cycle the pool cleanup is never installed. The patch you mentioned fixes that but it depends on an other interpreter management (the heart of the threading branch). Hence, if you need that use the threading branch. I use it for quite a time now in production with prefork. I know there are issues with worker. So, I don't recommend that. In modperl_callback.c at line 202 you'll find these lines: /* XXX: would like to do this in modperl_hook_create_request() * but modperl_interp_select() is what figures out if * PerlInterpScope eq handler, in which case we do not register * a cleanup. modperl_hook_create_request() is also currently always * run even if modperl isn't handling any part of the request */ modperl_config_req_cleanup_register(r, rcfg); This is inside the modperl_callback_run_handlers() function which is the central point that runs all PerlXXXHandlers. That's why you need at least one PerlXXXHandler to activate the PerlCleanupHandler. In the threading branch the modperl_config_req_cleanup_register() call could have been moved to the create_request hook as mentioned in the comment. Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: mod_perl survey results
On Mon 10 Nov 2008, Steven Siebert wrote: Let me know how to get involved How well is your C? There is a segfault waiting to be hunted down. It's one of the nicer. It happens each time the test suite runs with worker MPM. If interested I can give you further information. The threading branch at http://svn.apache.org/repos/asf/perl/modperl/branches/threading needs a few ifdefs to work with perls without threading support. That branch is also a good starting point for the segfault above. Try out the threading branch with the Windows MPM. I don't have Windows (and don't care much). So there may be a lot to do. Further on the line I have a few ideas how to make modperl consume less memory when running under a threaded MPM. I think we really need better threading support to survive. We can of course repeat the mantra use prefork ever again. But many apache developers tend to use a threaded MPM as default. Some days ago Paul Querna even asked how valuable copy-on-write is and if it can be dropped in favor of better Windows compatibility. The day will come when prefork is not the default MPM even on Unix-like systems. Of course there is always a need for advocacy, writing papers, presentations, case studies, cool solutions etc. Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: mod_perl survey results
On Mon 10 Nov 2008, André Warnier wrote: Ok guys, I'm nowhere as good a programmer as many people on this list, but a) I do have patience with beginners, b) I'm convinced and c) maybe I can do something in terms of documentation, if only to fix missing links. And d) I'd love to see my name somewhere as a contributor, even at the very end and in small font. So, where do I start ? c) See http://perl.apache.org/contribute/index.html. Lately there have been a few complains about missing, incomplete or misleading docs on the list. You could start to submit patches. A simple one could be this: http://www.gossamer-threads.com/lists/modperl/dev/98380. Find out why it was that way in the first place then patch it. Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: mod_perl survey results
On Mon 10 Nov 2008, Steven Siebert wrote: More memory but potentially faster, correct? Since we don't have to spawn as many processes to accommodate a load? Perl is a real memory hog. Byte-compiled code can become quite big. Multiply that with the number of perl interpreters running and you'll get huge numbers. With the prefork MPM we use the load all modules you need at startup mantra. This way the operating system comes to rescue with copy-on-write. The byte-compiled code is almost read-only after it is compiled. If it is compiled by the perl interpreter in the parent apache all that memory becomes shared memory. With a threaded MPM the situation is different. Here at startup we have one perl interpreter that loads all the modules. Then AFTER the child is forked it creates as many as configured additional interpreters. Each of those is a copy of the master interpreter. But now all the memory that holds the byte-compiled code is allocated again for each interpreter and almost nothing is shared by copy-on-write. So the creation of a new perl interpreter is really expensive. Although there is one sane usage of modperl with a threaded MPM, I believe. If your only modperl usage is to configure the request at runtime, do authentication etc you need the interpreter only for a small part of the request cycle. So you can configure say 3 interpreters to handle 250 or so threads. But for that to become usable we need a better modperl and perhaps better MPMs (event based). Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: Substituting URLs in PerlOutputFilterHandler
On Mon 10 Nov 2008, Puneet Lakhina wrote: essentially im trying to modify a href= tags in the outgoing HTML pages. Now I am trying to do this using PerlOutputFilterHandler. I have the following questions: a) Is this the right filter to use for an application like this? yes b) As per my understanding my hander can be invoked multiple times based on the no. of buckets. Is this correct? yes If yes, then how should I handle the situation where a part of the tag like say 'a href' comes in one bucket and '=' comes in another. The basic idea is to carry around the tail of the current buffer, the 'a href', in the filter context. The last handler in this chapter http://perl.apache.org/docs/2.0/user/handlers/filters.html#Stream_oriented_Output_Filters shows the technique. The OutputFilter routine in Apache2::ClickPath (on CPAN) is another example. Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: Fwd: MPM-safe mp2 Singleton Pattern?
On Fri 07 Nov 2008, Steven Siebert wrote:
Is there any way to have a MPM-safe (specifically thread-safe)
singleton pattern?
Modperl manages an interpreter pool if it runs under a threaded MPM.
That means the first modperl-related action in the request cycle pulls
an interpreter from the pool. Some time later the interpreter is the
put back to the pool. The point in time when that happens is defined by
the PerlInterpScope directive. It can happen immediately after the
current phase in the request cycle or when the request is over or when
the client connection is dropped. (And you can have a separate
interpreter for a subrequest.) Perl variables allocated in one
interpreter are accessible only from that interpreter unless they are
marked as shared. Shared variables are implemented by the
thread::shared module. This module creates another shared perl
interpreter that does the actual memory management. A shared variable
is then allocated inside that special interpreter and another one is
allocated inside the current interpreter. Then that variable is tied
to the variable in the shared interpreter in a similar way to what the
tie() command does.
So for what you want you need to make sure (if I understood your
problem) that the same instance is used throughout the request cycle.
Plus you want that it is destroyed when the request is over.
The simplest way is to put the object as a pnote:
sub getInstance {
my $r=shift;
return $r-pnotes-{Logger} if exists $r-pnotes-{Logger};
return $r-pnotes-{Logger}=Logger-new;
}
No global $this!
When the request pool is destroyed all pnotes are destroyed as well.
In theory that also binds the current interpreter to the pool. That
means it makes sure that the interpreter stays the same at least as
long as the request pool exists no matter what PerlInterpScope says.
But there were bugs with that in the past. So perhaps you have to
upgrade your modperl.
And finally an advice, don't use a threaded MPM with modperl in
production! If you want to help with the code it would be appreciated.
A good starting point is the threading branch at
http://svn.apache.org/repos/asf/perl/modperl/branches/threading
Torsten
--
Need professional mod_perl support?
Just hire me: [EMAIL PROTECTED]
Re: Mod Perl 2.0 HTTP Handlers
On Sun 02 Nov 2008, Torsten Foertsch wrote:
On Sun 02 Nov 2008, Puneet Lakhina wrote:
Input headers are found in $r-headers_in and are set even before
PerlPostReadRequest. So, have a close look at your code. If that
doesn't help try to create a test case as simple as you can and
post it to the list.
Not as per this:
http://perl.apache.org/docs/2.0/api/Apache2/RequestRec.html#C_heade
rs _in_
Then the documentation is wrong.
To show you that that is not an modperl-thing but implemented that way
in apache, let's have a look at the ultimate docs. In
httpd-2.x.y/server/protocol.c you'll find the ap_read_request function.
It reads in a request, finds the virtual host that handles the request
and calls PostReadRequest handlers. It is a bit lengthy. So I have
removed the unnecessary for the problem stuff and put in a few
comments:
request_rec *ap_read_request(conn_rec *conn)
{
...
r-headers_in = apr_table_make(r-pool, 25);
Here $r-headers_in is created as an empty table.
/* Get the request... */
if (!read_request_line(r, tmp_bb)) {
This reads the first request line, something like GET / HTTP/1.1
if (!r-assbackwards) {
ap_get_mime_headers_core(r, tmp_bb);
And this line reads in the headers and populates $r-headers_in.
if (apr_table_get(r-headers_in, Transfer-Encoding)
apr_table_get(r-headers_in, Content-Length)) {
You see, even the function itself uses $r-headers_in.
ap_update_vhost_from_headers(r);
Another example of $r-headers_in usage. Here the correct
NamedVirtualHost is found. This function needs the Host request header.
if ((access_status = ap_run_post_read_request(r))) {
And here a PerlPostReadRequestHandler is called.
...
}
Torsten
--
Need professional mod_perl support?
Just hire me: [EMAIL PROTECTED]
Re: Mod Perl 2.0 HTTP Handlers
On Sun 02 Nov 2008, Puneet Lakhina wrote: Initially I thought PerlTransHandler would be the right thing to use. It is. But as it turns out, the headers arent populated until that phase. Input headers are found in $r-headers_in and are set even before PerlPostReadRequest. So, have a close look at your code. If that doesn't help try to create a test case as simple as you can and post it to the list. Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: Mod Perl 2.0 HTTP Handlers
On Sun 02 Nov 2008, Puneet Lakhina wrote: One more question, is it possible for me to ensure that my module/or essenially mod_perl's processing of the URI translation happens before mod_rewrite or any other similar modules, and essentially mod_rewrite's rules work on the URI modified by me. That order is defined by the 4th parameter to ap_hook_translate_name(). mod_rewrite calls it as: ap_hook_translate_name(hook_uri2file, NULL, NULL, APR_HOOK_FIRST); and mod_perl as: ap_hook_trans(modperl_trans_handler, NULL, NULL, APR_HOOK_REALLY_FIRST); APR_HOOK_FIRST and APR_HOOK_REALLY_FIRST are integer numbers. Hence, you can write APR_HOOK_REALLY_FIRST-42 or so. The actual definitions can be found in srclib/apr-util/include/apr_hooks.h: /** run this hook first, before ANYTHING */ #define APR_HOOK_REALLY_FIRST (-10) /** run this hook first */ #define APR_HOOK_FIRST 0 /** run this hook somewhere */ #define APR_HOOK_MIDDLE 10 /** run this hook after every other hook which is defined*/ #define APR_HOOK_LAST 20 /** run this hook last, after EVERYTHING */ #define APR_HOOK_REALLY_LAST30 So, a PerlTransHandler is always called before the mod_rewrite translation handler since APR_HOOK_REALLY_FIRST is less than APR_HOOK_FIRST. The only way to change that is to patch and recompile one of the modules. Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: $r-filename, can I do this here ?
On Fri 17 Oct 2008, André Warnier wrote: In a PerlFixupHandler, I want to use $r-filename to reset the target of a PUT request, so as to trick the following Apache PUT content handler into writing the PUT-ted file somewhere else than what the original URL said. Yes you are allowed to do that. One thing you may have to consider is the $r-finfo object (sort of stat() result). It is created in the core maptostorage handler. So, if $r-filename is changed afterwards it is out of sync. Don't know if your response handler uses it, probably not. Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: $r-filename, can I do this here ?
On Fri 17 Oct 2008, André Warnier wrote: Any way I can kind of force Apache to update it's information after I call $r-filename, or does it already do that anyway ? See the example in http://perl.apache.org/docs/2.0/api/Apache2/RequestRec.html#C_filename_ Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: sethandlers question
On Tue 14 Oct 2008, Torsten Foertsch wrote:
On Tue 14 Oct 2008, André Warnier wrote:
When I try this, I get an Apache error with the following logfile
message : [Tue Oct 14 16:10:40 2008] [error] [client
84.158.163.207] $r-add_config() has failed: SetHandler not allowed
here at /usr/local/lib/apache2/perllib/AUTH/StarLogCookie.pm line
414.\n
That is an apache error. SetHandler needs the FileInfo override flag,
see http://httpd.apache.org/docs/2.2/mod/core.html#sethandler. So,
the correct command would be:
$r-add_config(['SetHandler ...'], Apache2::Const::OR_FILEINFO)
or simpler
$r-add_config(['SetHandler ...'], ~0)
(since you know what directive you are adding).
Maybe others are interested too, here are 2 private messages:
On Wed 15 Oct 2008, André Warnier wrote:
Ok, I did, and it seems to work too :
So I have 2 sequences that work :
a)
$r-add_config(['SetHandler modperl'], ~0);
$r-set_handlers(PerlAuthzHandler = []); # disable authorization
$r-set_handlers(PerlResponseHandler = \_send_login_form);
#$r-set_handlers(PerlFixupHandler = sub {
$_[0]-handler('modperl') } );
b)
#$r-add_config(['SetHandler modperl'], ~0);
$r-set_handlers(PerlAuthzHandler = []); # disable authorization
$r-set_handlers(PerlResponseHandler = \_send_login_form);
$r-set_handlers(PerlFixupHandler = sub {
$_[0]-handler('modperl') } );
Now, can you explain ?
I don't even understand what ~0 means..
On Wed 15 Oct 2008, Torsten Foertsch wrote:
On Wed 15 Oct 2008, André Warnier wrote:
1) you mentioned that it was due to a Apache bug.
Did I really say bug? I meant an Apache error message. You'll see the
same message if you try something like that:
Directory /
AllowOverride AuthConfig
/Directory
and then have an .htaccess with a SetHandler directive in it.
SetHandler needs AllowOverride FileInfo. But only AuthConfig is
set.
The 2nd parameter to add_config is a AllowOverride bit field. It
limits what kind of directives are allowed to be added. Think of it
as if you have read the directives from an untrusted source (like
.htaccess).
2) I do not have a clue why the above would work around that bug.
Apache wants here not only to be told to add the SetHandler but also
to be ensured Yes, that is what I want. Just do it!.
~0 means all bits turned on.
Torsten
--
Need professional mod_perl support?
Just hire me: [EMAIL PROTECTED]
Re: SetHandlers question, Phase II
On Wed 15 Oct 2008, André Warnier wrote:
I think you're wanting:
$r-set_handlers(PerlFixupHandler = sub
{$_[0]-handler('modperl')});
Yes. That works.
Better to return something sane from the handler:
sub {$_[0]-handler('modperl');0} # 0==Apache2::Const::OK
Also consider push_handlers instead of set_handlers. In case other
modules already have set a fixup handler.
Torsten
--
Need professional mod_perl support?
Just hire me: [EMAIL PROTECTED]
Re: Apache2::Request in PerlInitHandler and PerlAuthenHandler
On Tue 14 Oct 2008, Shibi NS wrote:
And the ARULink::MetalinkAuth::handler is
sub handler
{
my ($req) = @_;
my $status;
my $username;
my $userid;
my $email;
#
# Create a Session object.
#
my $req_params = $req-args;
my $cgi= CGI-new($req_params);
my $session= ARU::Session::get_cgi_session($cgi);
$req-subprocess_env;
}
When my handler reaches line '(my $cgi= new CGI($params);)' the
application is spinning ,seems like $req-args is returning null and
application line number 354 of CGI which is $req-subprocess_env;
Error from log file
Warning:
Deep recursion on subroutine CGI::new at
/m/isd/pm/ARULink/MetalinkAuth.pm line 114.
Don't know about CGI in this context but is there any chance to switch
to Apache2::Request (libapreq2)?
If you really see an empty $r-args at that stage while you are certain
to have passed something then it is an error in mod_perl. Try to create
a test case as short and simple as you can and file a bugreport.
Try something like
warn params: .$req-args;
The warning will appear in the error_log.
Can you please be more detailed about where it spins in CGI.pm? My
version 3.29 looks like:
350if ($MOD_PERL) {
351 if ($MOD_PERL == 1) {
352$self-r(Apache-request) unless $self-r;
353my $r = $self-r;
354$r-register_cleanup(\CGI::_reset_globals);
355 }
356 else {
357# XXX: once we have the new API
358# will do a real PerlOptions -SetupEnv check
359$self-r(Apache2::RequestUtil-request) unless $self-r;
360my $r = $self-r;
361$r-subprocess_env unless exists $ENV{REQUEST_METHOD};
362$r-pool-cleanup_register(\CGI::_reset_globals);
363 }
364 undef $NPH;
365}
Line 354 is somewhere in the mp1 code path. The mp2 code path looks
correct for me. Apache2::RequestUtil-request is set up in
postreadrequest and again in headerparser if +GlobalRequest.
cleanup_register() is also allowed here.
Ah, does your code check for a subrequest? Something like return ... if
$r-main or return ... unless $r-is_initial_req? If not you may see
it spinning in the subprocess_env line if your $r-path_info is not
empty. The reason is a subrequest issued during $r-subprocess in void
context to compute the standard CGI variable $ENV{PATH_TRANSLATED}.
Apache does that by $r-lookup_uri($r-path_info). Assuming you don't
have a download, ARULink or WebApp directory in your DocumentRoot the
whole URI will be in $r-path_info as well. So, the subrequest will hit
the same location container an thus the authen handler. Just a guess.
On Tue 14 Oct 2008, André Warnier wrote:
Now, below :
SetHandler perl-script
is used when your purpose is to run a traditional cgi-bin script
under mod_perl, not for embedded modules like your
PerlResponseHandler ARULink::handler
would indicate.
I suggest you try replacing this first by
SetHandler modperl
Sorry André but that is not entirely true. perl-script vs. modperl
simply means you want the more sophisticated response handler. It
affects only the response phase. Since the authen handler comes prior
to that it is irrelevant here. More sophisticated means you want %ENV
tied to $r-subprocess_env, tied *STDIN, *STDOUT, etc. The main reason
for me to avoid the perl-script handler is that it uses PerlOptions
+SetupEnv as default. That calls ap_add_cgi_vars() and that computes
$ENV{PATH_TRANSLATED} the IMHO most useless CGI variable. To do that it
issues a subrequest.
Torsten
--
Need professional mod_perl support?
Just hire me: [EMAIL PROTECTED]
Re: sethandlers question
On Tue 14 Oct 2008, André Warnier wrote: I doubly appreciate not only because my immediate problem is solved, but also because I learned a lot in the process. Try the $r-add_config again with the right override flag (OR_FILEINFO) and you'll see it will also work. It seems to me that this whole thread could be collected and re-arranged into a Tutorial on mod_perl phases, authentication, interactions with Apache etc.. but honestly at the moment I don't know where to begin. Find a place where you think it would fit and send in a patch, see http://perl.apache.org/contribute/svn_howto.html#Working_with_SVN on how to get the source. Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: sethandlers question
On Tue 14 Oct 2008, André Warnier wrote: When I try this, I get an Apache error with the following logfile message : [Tue Oct 14 16:10:40 2008] [error] [client 84.158.163.207] $r-add_config() has failed: SetHandler not allowed here at /usr/local/lib/apache2/perllib/AUTH/StarLogCookie.pm line 414.\n That is an apache error. SetHandler needs the FileInfo override flag, see http://httpd.apache.org/docs/2.2/mod/core.html#sethandler. So, the correct command would be: $r-add_config(['SetHandler ...'], Apache2::Const::OR_FILEINFO) or simpler $r-add_config(['SetHandler ...'], ~0) (since you know what directive you are adding). Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: sethandlers question
On Tue 14 Oct 2008, André Warnier wrote:
At this moment, I cannot see clearly how the ErrorDocument solution
would allow me to send back a login page created dynamically, as I
need to do. I will look it up, but as a definitive solution for the
time being I would prefer the $r-internal_redirect method, where I
can see more clearly how to do this.
Location /secured
PerlAuthenHandler sub { \
return Apache2::Const::FORBIDDEN unless ...;
}
ErrorDocument 403 /login
/Location
Location /login
SetHandler modperl
# you still have to supply a login_form function that expects a
# $r object.
PerlResponseHandler sub { \
my $r=shift; \
$r-content_type('text/html'); \
$r-print(login_form($r-prev));
return Apache2::Const::OK;
}
/Location
Your A-handler returns 403. Instead of sending the standard FORBIDDEN
page to the browser apache catches the error because of the
ErrorDocument directive. /login looks like an URI. Hence apache will
perform an internal redirect to it. /login is configured in the
normal modperl-way.
The browser will see an HTTP status of 403 (it will also be logged as
403). But the message sent along will be the output of login_form. And
that is what the user will see. login_form has access to the original
request. So, it can pass the original URI in a hidden field.
See also http://httpd.apache.org/docs/2.2/mod/core.html#errordocument
Instead of 403 you can also use 401 if you want to use
http-authentication. That is the user will see a browser generated
password box. But I doubt that is what you want.
But I still have some (hopefully) last questions, if you would be so
kind :
As I understand this, the internal_redirect() basically stops the
current request, and replaces it in-place by a totally new request,
except that in that new request, $r-prev will be set to the
(original) request that did the redirect, right ?
Yes. Best if you return OK from the current handler right after the
$r-internal_redirect line.
(So in the new request, I can still access the pnotes set by the
original request, via $r-prev-pnotes()).
Yes.
For this new request, the cycle restarts right at the beginning, and
runs through all the steps just as if this was a main request. Right
?
AAA are possibly skipped as I mentioned before.
Are there any other significant differences between this new request
and the previous one that I should be very aware of (in a
mod_perl/AAA context) ?
You'll find some, ;-)
One thing I do not really understand, is what happens to the original
request, when it issues a call to $r-internal_redirect().
Does the original request still run to completion, or does it abort
right there ?
I think you'll see the original request again in the logging phase.
Does Apache still return something to the browser for the original
request, or is that completely replaced by whatever the new
(redirected) request produces ?
No, not unless you have sent something prior to the internal_redirect.
After the call to $r-internal_redirect(), should the original
request still do something ? (like return Apache2::Const::?) ?
OK, see
http://perl.apache.org/docs/2.0/api/Apache2/SubRequest.html#C_internal_redirect_
P.S.
I have downloaded the Apache source, and looked at the code within
httpd-2.x.y/server/request.c, but it is still a bit obscure to me,
except in a very general sense. I kind of approximately follow the
steps, but do not always understand aspects like why it re-does
directory walks at several stages, nor can I see at all where
mod_perl plugs in to that code.
great! Many treasures are hidden there.
Torsten
--
Need professional mod_perl support?
Just hire me: [EMAIL PROTECTED]
Re: sethandlers question
On Tue 14 Oct 2008, André Warnier wrote:
reminder of my config :
LocationMatch /servlet.xyz$
SetHandler jakarta-servlet
...
/Location
And, in summary, what I was trying to do is :
if the PerlAuthenHandler finds that the request is not authenticated,
then it tries to change the response handler for this request, from
the original jakarta-servlet handler, to my own response handler,
which in fact returns a login page.
(I need something active to return the login page, because I need
to insert some data into the login page prior to returning it).
And for that I used this code in the PerlAuthenHandler :
# set our own response handler
$r-handler('modperl');
$r-set_handlers(PerlResponseHandler = \_send_login_form);
return OK;
The problem is the $r-handler statement. The program flow is this:
- ...
- uri translation
- map to storage
- Location container is applied
the SetHandler directive is applied to the requests configuration
structure. That does not mean $r-handler is set. It is simply noted
there is a SetHandler directive active for the request.
- aaa
- ...
- fixup
the SetHandler value is copied to $r-handler. See
httpd-2.x.y/server/core.c:core_override_type()
if (conf-handler strcmp(conf-handler, none))
r-handler = conf-handler;
That means $r-handler is set only if there was a SetHandler directive
and if it was not SetHandler none.
So, that leads us to 4 possible solutions:
1) remove the SetHandler directive from the Location container and set
$r-handler either to jakarta-servelet or to modperl in your aaa
handler. It won't be touched if other modules don't kick in. mod_dir
comes to mind but it reads (fixup handler):
/* In case mod_mime wasn't present, and no handler was assigned. */
if (!r-handler) {
r-handler = DIR_MAGIC_TYPE;
}
So, that's not a problem. But there may be other modules that don't
check that.
2) replace $r-handler('modperl') in the aaa phase by
$r-add_config(['SetHandler modperl']). This way you override the
handler set in the configuration structure. In the fixup phase it is
then automagically copied to $r-handler.
3) leave the $r-handler('modperl') and add a
$r-add_config(['SetHandler none']);
4) set $r-handler in a fixup handler.
$r-add_config works very similar to
Location /
your directives
/Location
The only difference is the time when it is applied to the config struct.
That config is reset after the translation phase. So it is useless to
try $r-add_config before. Location containers from the httpd.conf are
applied after MapToStorage. So anything you do in your own MapToStorage
handler can be overridden in Location containers even if your handler
returns OK to skip the core MapToStorage handler.
But anything applied after that, that means from the header-parser phase
onward overrides the httpd.conf settings.
One of the possibilities envisioned previously was :
Or, a thought that just occurred to me, is the solution for me
(stopping jakarta-servlet from running and running my
PerlresponseHandler instead), as easy as setting the no-jk
variable (if I can do that within my PerlAuthenHandler, and how) ?
I was tempted to suggest that. But I don't know nothing about
jakarta. If you want to try use $r-subprocess_env-{'no-jk'}=1;
I tried this, because it was the smallest possible change to my
module. It works, in the sense that it sets the no-jk variable,
which consequently has the effect of making the jakarta-servlet
handler above not process this link.
But it is not a solution for me, since what happens is that the
jakarta-servlet handler is in fact still executed by Apache as the
response handler, but it probably just returns DECLINED.
That's what I thought it would.
Torsten
--
Need professional mod_perl support?
Just hire me: [EMAIL PROTECTED]
Re: sethandlers question
On Sun 12 Oct 2008, André Warnier wrote:
Location /some/url
SetHandler jakarta-servlet
SetEnvIf REQUEST_URI \.(htm|web|css|gif|jpg|js|html?)$ no-jk
PerlXXXHandler My::Module-some_method
...
/Location
(jakarta-servlet above means mod_jk -- Tomcat)
(and PerlXXXHandler being any kind of Perl HTTP Handler running soon
enough)
The Question :
Is it possible, in the PerlXXXHandler, on a request-by-request base,
to disable the jakarta-servlet handler, and install the following
instead
SetHandler modperl
PerlResponseHandler My::Module-some_other_method
?
I know that I can do something like the above for regular static
pages, and I use this already :
$r-handler('modperl');
$r-set_handlers(PerlAuthzHandler = []); # stop authorization
from running
$r-set_handlers(PerlResponseHandler = \_send_login_form);
However, when I try to do the same in a Location like the above, it
seems that the jakarta-servlet handler runs anyway, and my
PerlResponseHandler is never called.
Does that piece of code run in a /Perl(PostReadRequest|Trans|
MapToStorage)Handler/ ?
Please, have a look again at the ap_process_request_internal() function
in httpd-2.x.y/server/request.c. This function implements almost the
complete request cycle (save PostReadRequest, Response and Log).
Just after ap_run_translate_name(), that is the PerlTransHandler, you'll
see this piece:
/* Reset to the server default config prior to running
map_to_storage
*/
r-per_dir_config = r-server-lookup_defaults;
if ((access_status = ap_run_map_to_storage(r))) {
/* This request wasn't in storage (e.g. TRACE) */
return access_status;
}
/* Rerun the location walk, which overrides any map_to_storage
config.
*/
if ((access_status = ap_location_walk(r))) {
return access_status;
}
/* Only on the main request! */
if (r-main == NULL) {
if ((access_status = ap_run_header_parser(r))) {
return access_status;
}
}
...
The line r-per_dir_config = r-server-lookup_defaults; resets all
request specific configurations applied so far. After that line the
request is configured as if no Directory/Location or other container
were present in your httpd.conf. Next comes MapToStorage. Here
Directory, DirectoryMatch, Files and FilesMatch containers are applied
to the request. Then, and this is what I think hits you, comes an
ap_location_walk. Here Location containers are applied. So, your
SetHandler inside the Location overrides any $r-handler seen before.
After the ap_location_walk comes the PerlHeaderParserHandler. But it is
skipped for subrequests. A further examination will show you that the
next phase that is run by all requests is the PerlTypeHandler just
before Fixup. So, move your $r-handler('modperl'); to one of those
and it will work. I prefer Fixup but that is based only on the fact
that it is documented as do-what-you-want phase. There are directives
that change the handler in the type-checker phase. In fact the
AddHandler directive works that way. Mod_mime looks at the file
extension and adjusts the handler, e.g AddHandler cgi-script .pl. So
using Fixup your module is the last to kick in.
To carry state around use pnotes, notes or subprocess_env.
Torsten
--
Need professional mod_perl support?
Just hire me: [EMAIL PROTECTED]
Re: sethandlers question
On Mon 13 Oct 2008, André Warnier wrote:
As explained the other thread (requests and subrequests), I thought
it was more elegant and mod_perl-ish, instead of doing a re-direct to
a login page if the user needed authentifying, to having the login
page returned by a content handler.
The apache-way to do this has a name: ErrorDocument or in mod_perl
$r-custom_response. Configure an ErrorDocument for 403 and return
HTTP_FORBIDDEN in your A...Handler. That has the benefit that the 403
status is logged and that the browser (think about search engines) sees
it. But the custom response page must be larger than 512 bytes or
you'll encounter problems with IE. If you don't care about that you can
use $r-internal_redirect. That is a bit less work to do for your
server.
Please, have a look again at the ap_process_request_internal()
function in httpd-2.x.y/server/request.c.
I would first have to know where this thing is. I have never so far
looked at the Apache source code, so I need some pointers there.
Not that if I look, I would necessarily understand much of it, but I
can try.
Now, why would I, since you provide the nice explanation below ?
Out of curiosity! Just download the apache source untar it and open in
in your favorite editor. C is not that much different from perl and the
request.c file is really easy to read. You don't have to modify it,
just try to understand what is going on!
So, you mean that for every and each request, Apache re-configures
itself ? If that is so, it is quite amazing the speed at which it
can serve a page..
Just keep in mind for the following that apache is originally designed
to have a static configuration. So at start it splits the configuration
into blocks, one holds the basic config that is outside all containers
and others for each Location/Directory/etc. From those blocks it
generates pre-parsed configuration vectors. At runtime each request
starts with the basic config fetched from r-server-lookup_defaults
then it looks for other vectors matching the request and merges them.
But there is another configuration source, .htaccess. Those are read in,
compiled and merged in the same function (ap_directory_walk() (hard to
understand)) that applies Directory containers (called from
ap_run_map_to_storage()).
With mod_perl comes yet another configuration source, $r-add_config.
This is implemented very similar to .htaccess only the directives are
not read from a file but from memory and they are compiled and merged
at the time $r-add_config is called.
Well, I am not sure :
The handler set in the configuration file for that Location, is
originally jakarta-servlet (as per the snippet below).
It is I, in my PerlAuthenHandler, who is trying to reset this to
modperl, and in mod_perl, to my send_login_page() method.
Before I do that, there is no $r-handler nor PerlResponseHandler
set.
So, why not issue a $r-internal_redirect. I don't exactly know when the
SetHandler directive sets $r-handler. Maybe it is done in the type
checker phase. So perhaps you have to move your code to Fixup or use
$r-add_config(['SetHandler modperl']).
I don't think I can do that, unless I want the PerlTypeHandler or
PerlFixupHandler to become an authentication handler. That would
probably for one upset Adam, for two be kind of funky, no ?
Why not?
if( $override_handler ) {
$r-push_handlers( PerlFixupHandler=__PACKAGE__.'::fixup' );
$r-pnotes-{override_handler}='modperl';
}
...
sub fixup {
$_[0]-handler($_[0]-pnotes-{override_handler};
return Apache2::Const::OK;
}
Now, in summary :
This is a very specialised authentication module, meant to be used in
very specific circumstances and not for general usage. I control the
environment, so I can pretty much do what I want as long as it does
not crash the system or present some big security hole or the like.
In my case, that Location containing the
sethandler jakarta-servlet
is pretty much unique: it is not a sub- or super-location of anything
else, and it fulfills only one very narrow purpose.
To my knowledge, there will be no other module playing a role there,
other than the jakarta-servlet initial handler, plus whatever I
decide to put in there as mod_perl modules and methods.
This for the entire duration of one request.
Now, assuming I have no other stuff in there than the following lines
LocationMatch /servlet.xyz$
SetHandler jakarta-servlet
SetEnvIf REQUEST_URI \.(htm|web|css|gif|jpg|js|html?)$ no-jk
AuthType MyModule
AuthName MyXyzApp
PerlSetVar MyModule_var1 value
PerlAccessHandler My::Module-some_access_method
PerlAuthenHandler My::Module-some_auth_method
PerlAuthzHandler My::Module-some_authz_method
require valid-user
/Location
then, if I understand correctly what you write above, since I have no
PerlTransHandler, nor Fixup, etc.. I should be able to do the same
in the PerlAuthenHandler, is it no so ?
Or
Re: requests and sub-requests
On Sun 12 Oct 2008, André Warnier wrote:
In an attempt at being clever, I put the following code in the
handler :
unless ($r-is_initial_req) {
if (defined $r-prev) {
# we are in a subrequest. Just copy user from main
request. $r-user( $r-prev-user );
}
# Also disable authorization phase
$r-set_handlers(PerlAuthzHandler = undef);
return OK;
}
You have to distinguish between subrequests and internal redirects. The
former result from $r-lookup_uri, $r-lookup_file or similar (there
are a few more such functions in the C API) and internal redirects that
result from $r-internal_redirect (internal_fast_redirect() is not as
the name suggests an internal redirect but simply overrides the current
request). Subrequests are used for example by mod_rewrite, mod_include,
mod_negotiation to look for some characteristics of a document and
perhaps pull it in (run() it). Internal redirects are used in mod_cgi
when the CGI output indicates a status 200 (HTTP_OK) but also contains
a Location header. But the main usage of internal redirects is the
ErrorDocument.
Now, is_initial_req() checks if the current $r is the result of a
subrequest or the result of a internal redirect and returns false if
so. prev() returns the parent request if the current $r is the result
of an internal redirect and main() returns the main request if the
current $r is a subrequest. So, your code checks only for internal
redirects (ErrorDocument).
Now, have a look at httpd-2.x.y/server/request.c around line 170. You'll
see this piece of code:
/* Skip authn/authz if the parent or prior request passed the
* authn/authz,
* and that configuration didn't change (this requires
* optimized _walk()
* functions in map_to_storage that use the same merge results given
* identical input.) If the config changes, we must re-auth.
*/
if (r-main (r-main-per_dir_config == r-per_dir_config)) {
r-user = r-main-user;
r-ap_auth_type = r-main-ap_auth_type;
}
else if (r-prev (r-prev-per_dir_config == r-per_dir_config))
{
r-user = r-prev-user;
r-ap_auth_type = r-prev-ap_auth_type;
}
else {
switch (ap_satisfies(r)) {
case SATISFY_ALL:
case SATISFY_NOSPEC:
if ((access_status = ap_run_access_checker(r)) != 0) {
return decl_die(access_status, check access, r);
...
You see, you are not the first who had had the idea of reusing an
established identity. If your subreq or internal redirect hits the same
Location or Directory container the AAA phases are completely skipped.
Maybe this is enough optimization if you shift a few directives around
in your httpd.conf.
If not, the code above shows you how to do it. But you must ask yourself
if it really is valid to reuse the identity. I believe, you can safely
inherit the identity from $r-main or $r-prev but you must not skip
the other 2 A's. If you can't it would mean you have one realm of
identities for the main request and another for the subreq. That, I'd
say, is a configuration error.
The idea being that if we are in a sub-request, there is no point in
authenticating/authorizing it again, since the main request should
already do that, right ? Optimisation..
Now the above works very nicely, except in the case where, before
this handler gets called, there is an intervention by mod_rewrite. It
seems as if mod_rewrite makes the above fail, even when the rewrite
condition does not apply and the URL is considered as a
pass-through.
I suspect that it is because mod_rewrite, no matter what, invoques
the original (or modified) URL as a sub-request of the original
request. This would cause the above to fail, because in such a case,
the above conditional code would be invoked, but there is no
$r-prev-user to be copied.
mod_rewrite doesn't make subrequests if not asked to. I know only of 2
ways to have mod_rewrite perform a subreq: %{LA-U:variable}
and %{LA-F:variable} in a RewriteCond.
Torsten
--
Need professional mod_perl support?
Just hire me: [EMAIL PROTECTED]
Re: requests and sub-requests
On Sun 12 Oct 2008, André Warnier wrote: I have a little question related to the above, but not very urgent : why the check on the configuration change ? what can change between a request and a sub-request (or internal redirect) ? Suppose this: Location /mainreq Require group foo /Location Location /subreq Require group bar /Location If during the processing of /mainreq a subreq is issued to /subreq the other Require group must be respected. Of course there also can be another AuthUserFile or so, so that the identity could not be established and the subreq results in a 401. I'd consider that as a configuration error. The configuration change mentioned above means the different Location containers. They are represented as $r-per_dir_config. I don't think so, because this is a really specific authentication method, for a special case. And I don't think that Apache will skip the mod_perl AAA phases, will it ? Yes, mod_perl handlers are called inside these ap_run_... functions. ap_run_access_checker() is the first of the 3 A's. A PerlAccessHandler is called from this function. Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: questions on Apache2::Connection
On Sat 11 Oct 2008, André Warnier wrote:
Do I understand this correctly that if in a Perl Handler i get
my $c = $r-connection();
then $c is the object that represents the persistent TCP/IP
connection between the browser and the server, in case there is
keep-alive going on ?
Now lets say that I create an authentication method based on the
Request (as they tend to be usually).
On the first request, the authentication happens, and I set a
$c-notes('credentials') value. I also set a browser cookie.
Yes, connection notes and connection pnotes are persistent across
keep-alive requests.
On subsequent requests, I could check this $c-notes('credentials')
first, in case a previous request over the same connection already
resulted in authentication, could I not ?
In the worst case, the connection is new and I would not have these
notes (meaning I then need to get the cookie, and in its absence redo
an authentication); but in the vast majority of cases (depending on
keep-alive), I could save myself some overhead by considering the
connection as authenticated instead of the request, no ?
Or are there some pitfalls here of which I am ignorant ?
Or is the potential gain not worth the cost of getting the
$r-connection ?
I see 2 points to consider:
1) A reverse proxy in front of the web server can maintain a persistent
connection to the backend but server different clients and thus spoil
your caching.
2) The combination of prefork-MPM, mod_perl and keep-alive is perilous
on the Internet because one apache process is locked over the whole
keep-alive time. A malicious client sends one request and let the kept
alive connection time out by the server. A single client can eat up all
your servers in a very short time. Of course a similar attack is
possible based on the server's TimeOut setting but they are a bit
trickier. You deploy that combination directly on the Internet but you
have to have a close look at the TimeOut and KeepAliveTimeout settings.
Make them as short as you can.
Torsten
--
Need professional mod_perl support?
Just hire me: [EMAIL PROTECTED]
Re: Free to wrong pool error in Apache logs under high load
On Fri 10 Oct 2008, Scott Tomilson wrote: I am running a build of mod_perl I got via perl.apache.org, yes (or I think probably linked to uwinnipeg?). No other modules than the standard ones, plus the standard ActivePerl download from ActiveState. We actually tried the same solution on Solaris and it worked fine. It does seem specific to Windows itself - or more likely ActivePerl on Windows... The windows MPM is multithreaded, the prefork MPM that you have probably used on Solaris is not. That makes a difference. Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: Internal redirect inside an input filter
On Thu 09 Oct 2008, Dan DeSmet wrote: I'm attempting to write an input filter that performs an internal redirect based on the contents of the cookies sent in the request headers. Why an input filter? What you want is better done in a PerlTransHandler or a PerlFixupHandler. Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: [MP2]: strange behavior with Apache2::SubRequest::run
On Fri 19 Sep 2008, titetluc titetluc wrote: Does your mod_perl one return Apache2::Const::REDIRECT at the end? No, the module returns Apache2::Const::MOVED_TEMPORARILY, setting the Location header by using $r-err_headers_out $ perl -MApache2::Const=REDIRECT,HTTP_MOVED_TEMPORARILY -le 'print REDIRECT; print HTTP_MOVED_TEMPORARILY' 302 302 Guess what that means. Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: How to extract the protocol part of URL from an apache request
On Wed 17 Sep 2008, grsvarma019 wrote: But , i couldn't find how to extract the protocol(http or https ) There are Apache2::ModSSL and Apache::SSLLookup on CPAN in case you need that information in a request phase prior to the ResponseHandler. Mod_ssl can be configured to export SSL information as environment variables. Those can be looked up in a ResponseHandler via $r-subprocess_env. Further, if you are unable to install an XS module (precompiled mod_perl+apache on windows without C compiler for instance) and you need SSL information prior to the response phase you can issue a subrequest to get it. Not the fastest way but it works. I have described that technique in a previous mail to the list this or last year. Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: How to extract the protocol part of URL from an apache request
On Wed 17 Sep 2008, John ORourke wrote:
I had the same problem but the machine serving the request had a
reverse proxy in front of it. I used the following to inject a
header on the proxy:
SetEnv SCHEME http
RewriteCond %{HTTPS} on
RewriteRule ^(.*) $1 [E=SCHEME:https]
Yes, I think that should work as well without the reverse proxy.
Mod_rewrite talks directly to mod_ssl.
Torsten
--
Need professional mod_perl support?
Just hire me: [EMAIL PROTECTED]
Re: How to pass the post data to a script
On Wed 17 Sep 2008, grsvarma019 wrote:
If the requested script of an application needs POST data, that data
is not being sent to the requested script ,but instead it is being
captured by apache request object of the mod_perl.
How can i pass this data to the requested script?
First, see if libapreq2 can help!
If it doesn't you can try to implement a request input filter that
caches the data on disk while mod_perl does $r-read or
$r-discard_request_body and a second one that reinserts that input for
the script. Not sure if that works.
The first part looks similar to that:
use Apache2::Const -compile=qw/M_POST OK/;
use APR::Const -compile=qw/SUCCESS/;
use Apache2::Filter ();
use APR::Bucket ();
use APR::Brigade ();
if( $r-method_number==Apache2::Const::M_POST ) {
my @content;
my $cl=0;
my $rc=$r-add_input_filter( sub {
my ($f, $bb, $mode, $block, $readbytes) = @_;
my $rv = $f-next-get_brigade($bb, $mode, $block, $readbytes);
return $rv unless $rv == APR::Const::SUCCESS;
for (my $b = $bb-first; $b; $b = $bb-next($b)) {
$b-read(my $bdata);
$cl+=length $bdata;
push @content, $bdata;
}
return Apache2::Const::OK;
} );
$r-discard_request_body;
$r-pnotes-[EMAIL PROTECTED];
$r-pnotes-{rbody_length}=$cl;
}
Well, it saves the request as pnotes but it shows the principle. The
second part is a bit trickier if not impossible. You'll have to modify
the input filter chain again, insert a filter that creates buckets from
the saved data. But I don't know if that works once EOF has been seen
on input.
Torsten
--
Need professional mod_perl support?
Just hire me: [EMAIL PROTECTED]
Re: using MOD_PERL with script causes looping
On Wed 10 Sep 2008, csross wrote: On a solaris 8 server with Apache 2.6 and perl 5.8.8, we have an html/javascript file that uses HTTP Post to a perl script. The perl script sends data back via HTTP get to a redirected HTML file. The HTML file strips off data from the command line of the URL (suff after the ?) and then posts more data to the perl script until all of the data it needs is retreived. This works fine in perl without mod-perl enabled, but loops in modperl, as if the initial connection for the first request is still open and the same data is being endlessly sent back. This happens on later versions of Solaris as well. We did find that after what seems like many interations of the looping, the data finally get processed and the subsequent screen appears, but that is after a minute at least. When not using mod_perl, this doesn't happen. The developer wanted to put in a wait type command to see if that worked, but it did not make a difference. Has anyone seen this type of behavior? I think the reason is that the perl script holds some status. The difference between mod_perl and CGI is that the former uses the same interpreter again and again while the latter starts the interpreter anew each time. So, if the script doesn't clean up things properly it may look different for the second request. You probably also use KeepAlive requests. That would explain the fact that it worked after many requests. Either the server or the browser has dropped the connection. The browser then establishes a new one and probably hits another perl interpreter. That one is still in pristine state. Hence it worked. Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: mod_perl 2 and IPC::Open3
On Thu 04 Sep 2008, Tina Müller wrote:
$pid = open3($wtr, $rdr, $err,
'/usr/bin/source-highlight', '-s', 'perl', '-css',
'--no-doc');
print $wtr $content;
close $wtr;
warn __PACKAGE__.':'.__LINE__.: before read loop\n;
while ($rdr) {
# this is not executed
warn __PACKAGE__.':'.__LINE__.: line $_\n;
$highlighted .= $_;
}
Don't know if it relates to the problem but this code is quite fragile.
It depends upon whether $content completely fits into the operating
systems pipe buffer. In your program the print $wtr or the close $wtr
statements may infinitely block if $content is too large.
To illustrate that try this:
perl -e 'pipe my ($r, $w) or die pipe: $!\n; print $w xx64000;
warn printed\n; close $w;'
Here a pipe is created and written to but never read from. This example
succeeds on my linux box. If 64000 is raised to 69000 the print $w
succeeds and I see the output of warn(). But the process never ends. It
is blocked in the close() operation. If the buffer length is raised a
bit further to 7 I don't even see the warning. The process is
blocked in the print() operation. The pipe buffer size depends upon the
operating system. So your numbers may vary.
I'd recommend to rewrite that piece based on either perl's select(),
IO::Select or some kind of event library like Lib::Event, EV, Event or
even POE. In all cases you have to use non-blocking IO for reading and
writing combined with sysread/syswrite.
Further, I'd suggest a bit of error handling. All those operations
open3, print, close may fail. Only open3 reports the failure via an
exception.
Also, you may want to watch out for SIGPIPE.
Torsten
--
Need professional mod_perl support?
Just hire me: [EMAIL PROTECTED]
Re: question on downloading CGI
On Fri 05 Sep 2008, [EMAIL PROTECTED] wrote:
I want to generate a data file which should be downloaded by clients.
Rather than generate this file and put it in a web dir and tell
clients to download it, is there any way to generate the content
dynamicly and put it to cients? I mean I don't want to generate the
temporary file.
Plenty.
For example in your httpd.conf
Location /download
SetHandler modperl
PerlResponseHandler sub { \
use Apache2::RequestRec ();\
use Apache2::RequestIO (); \
use Apache2::Connection ();\
$_[0]-content_type('application/octet-stream'); \
until($_[0]-connection-aborted) {\
$_[0]-print(rand); \
} \
return 0; \
}
/Location
Now restart your server and then
curl -v http://server/download
And you'll get never ending randomness.
You can also write a little shell script and use it via mod_cgi:
the shell script
#!/bin/bash
echo Content-Type: application/octet-stream
echo
cat /dev/urandom
end of shell script
Now fetch the Apache docs and configure your server to execute the
script as CGI program.
Remember it is not necessarily faster to send dynamic content this way.
Sometimes it is faster to write a temporary file sometime before the
response phase and let the default handler send it. If you decide to
send the content on the fly try to figure out the content length and
set the Content-Length header. This can also improve your performance.
Torsten
--
Need professional mod_perl support?
Just hire me: [EMAIL PROTECTED]
Re: Prototype mismatch: sub main::strftime
On Tue 26 Aug 2008, Darragh Gammell wrote: Subroutine main::strftime redefined at /usr/share/perl/5.8/Exporter.pm line 66. at /home/secure/public_html/tagtag/send.cgi line 10 Prototype mismatch: sub main::strftime: none vs ($\@;$) at /usr/share/perl/5.8/Exporter.pm line 66. at /home/secure/public_html/tagtag/send.cgi line 10 I think use Date::Format (at line 10 in send.cgi) exports its own version of strftime. That way it redefines the one exported by POSIX. That leads to the redefined warning. Further, one of the functions has a prototype while the other one hasn't. Hence comes the prototype mismatch. Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: mp2, IPC::Run3 Environment Variables giving variable not set error
On Sat 16 Aug 2008, Berg, Eric wrote: Now, the environment is a process global resource. So, if those values are changed all threads are affected. This is surely no what you want. When you say that the environment is a global process, global to what? Each forked process has its own environment, doesn't it? It seems to me from the brief piece in the User Guide about this and what you've said, that threaded environments are the real problem here. In the case of threaded environments, the environment is shared among all threads, whereas forked processes can and do have distinct environments per process. Is this correct? That means (as suggested in the relevant source comments) that a solution for the forked MPM would be much easier than for those that are threaded. Right? Correct. The words a process global resource above mean a resource that is global to the process. Btw, another way to manipulate the environment is using the modperl handler instead of perl-script. This way %ENV remains tied to the environment and changes will show up after a perl fork. See also http://perl.apache.org/docs/2.0/user/troubleshooting/troubleshooting.html#C_Libraries_Don_t_See_C__ENV__Entries_Set_by_Perl_Code Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: mp2, IPC::Run3 Environment Variables giving variable not set error
On Fri 15 Aug 2008, Berg, Eric wrote: This issue of the environment variables' not being passed to forked processes is turning into a fairly substantial problem for me. We have a number of places in our code -- both the modules that are tightly controlled and fairly easy to manage, as well as a bunch of scripts and CGI's that use them -- that use system or backticks (sometimes via IPC::Run3) to get info from external tools. Does anybody know anything about this issue or about the code involved? How difficult would it be to address this? I took a look at the code in modperl_env.c as suggested in the docs, but I'm a Perl guy, not a C guy. This is a deal killer for us. Let's have a look at the reason for this behavior first. It's always better to know what you are doing. Apache 2.x can use various MPMs. Some of them use threads instead of processes to achieve parallelism. Now, the environment is a process global resource. So, if those values are changed all threads are affected. This is surely no what you want. Therefore there is a request specific data structure called subprocess_env where environment variables for CGI scripts are kept. If you use the Apache interface to spawn subprocesses (Apache2::SubProcess) you'll notice that variables set in subprocess_env appear as environment variables of the spawned process. Under mod_perl %ENV is tied to $r-subprocess_env. If you are using the prefork MPM (as you probably do) you can still manipulate the environment without affecting other requests. But you have to use the C-level functions. There is a module called Env::C on CPAN that provides a perl interface to them. Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
Re: mod_perl 2.0 fails on make
On Fri 15 Aug 2008, Fred Moyer wrote: When I run 'make'. It goes pretty fair but it ends with this. modperl_exports.c:1169: error: `modperl_thx_interp_get' undeclared here (not in a function) modperl_exports.c:1173: error: `modperl_thx_interp_set' undeclared here (not in a function) make[1]: *** [modperl_exports.o] Error 1 make[1]: Leaving directory `/home/charlier/modperl-2.0/src/modules/perl' make: *** [modperl_lib] Error 2 Is there any known fix to work around this? Try with the 2.0.4 stable version of mod_perl. Also I would suggest building mod_perl as a shared object (dynamically, not statically), so would say remove the MP_USE_STATIC=1. I've seen people get it to work but almost everyone I know builds mod_perl as a loadable module instead of into the httpd binary. Agreed, the static build is a bit neglected. But without MP_USE_STATIC also the current SVN version should work. Torsten -- Need professional mod_perl support? Just hire me: [EMAIL PROTECTED]
