On 4/14/03 6:11 PM, Iain Truskett wrote:
> * John Siracusa (siracusa@...) [15 Apr 2003 07:06]:
>
> [...]
>> This is troubling...is the fragment only a client-side feature?
>
> Yes.

Ick.  I guess that means it's impossible to redirect someone to a secure
version of the exact URL that was requested, because I have no way of
knowing what the fragment part of the URL is!  Who designed this protocol
anyway... ;)

-John

John Siracusa wrote:
> On 4/14/03 6:11 PM, Iain Truskett wrote:
>
>>* John Siracusa (siracusa@...) [15 Apr 2003 07:06]:
>>
>>[...]
>>
>>>This is troubling...is the fragment only a client-side feature?
>>
>>Yes.
>
>
> Ick.  I guess that means it's impossible to redirect someone to a secure
> version of the exact URL that was requested, because I have no way of
> knowing what the fragment part of the URL is!  Who designed this protocol
> anyway... ;)

As a workaround you could do:

   /foo/bar/blah#blah or /foo/bar?blah#blah

where /foo/bar is the real resource, of course you will have to modify your
HTML to have the #fragment also included at the end of the url.

__________________________________________________________________
Stas Bekman            JAm_pH ------> Just Another mod_perl Hacker
http://stason.org/     mod_perl Guide ---> http://perl.apache.org
mailto:stas@... http://use.perl.org http://apacheweek.com
http://modperlbook.org http://apache.org   http://ticketmaster.com

Jozef Nowak wrote:
> Hello, all.
>
> Is it possible to check the version of modperl from the httpd.conf file? I
have the version check in my filter, but would prefer checking the proper
modperl version upon Apache startup.

You mean mp1 vs mp2? or 1.27 vs 1.26?

The former is documented here:
http://perl.apache.org/docs/2.0/user/porting/porting.html#Making_Code_Conditiona\
l_on_Running_mod_perl_Version

The latter mechanism doesn't exist, but can be added.

__________________________________________________________________
Stas Bekman            JAm_pH ------> Just Another mod_perl Hacker
http://stason.org/     mod_perl Guide ---> http://perl.apache.org
mailto:stas@... http://use.perl.org http://apacheweek.com
http://modperlbook.org http://apache.org   http://ticketmaster.com

Jim Brandt wrote:
> I did some additional testing on Mac OS X with Perl 5.8. I didn't see
> the Seg faults, but I did notice something that was the same as on Solaris.
>
> When running under Apache::PerlRun (these problems are all under
> Apache::PerlRun) on both systems, I also get "subroutine xxx redefined"
> warnings in the error log for scripts with even the simplest
> subroutines. These seem to appear as each child servers its second
> request. This happens when the sub in in the bottom of the main program
> or require'd from a separate file.
>
> I added:
>
>  PerlSetVar PerlRunOnce On
>
> and the "subroutine redefined" messages went away, but with a massive
> performance hit.
>
> Is it possible that I set up two faulty mod_perl configurations such
> that PerlRun isn't flushing the namespaces properly? If this is the
> case, could it be that the memory errors are a side effect of this?

Rather than guessing what could be the problem, why not get a core file and
look at its trace?

> On Friday, April 11, 2003, at 01:25  PM, Jim Brandt wrote:
>
>> We're running mod_perl with apache on Solaris 8 with Perl 5.6.1. While
>> doing some load tests on a script, we've seen intermittent seg faults
>> and bus errors. These only appear under load:
>>
>> [Thu Apr 10 15:09:06 2003] [notice] child pid 11276 exit signal
>> Segmentation Fault (11)
>> [Thu Apr 10 15:09:49 2003] [notice] child pid 11265 exit signal Bus
>> Error (10)
>
>
>
> ==========================================
> Jim Brandt
> Administrative Computing Services
> University at Buffalo


--


__________________________________________________________________
Stas Bekman            JAm_pH ------> Just Another mod_perl Hacker
http://stason.org/     mod_perl Guide ---> http://perl.apache.org
mailto:stas@... http://use.perl.org http://apacheweek.com
http://modperlbook.org http://apache.org   http://ticketmaster.com

Sorry for the off-topicness.  I am looking for a mailing list on
embedded Perl (like mod_perl is an embedded Perl interpreter in Apache,
   not like EmbPerl) and can't find anything.  Does anyone know is such a
list exists?

Please write me privately so we don't go off-topic on this list.

Thanks!

-- Brett
                                            http://www.chapelperilous.net/
------------------------------------------------------------------------
You will be reincarnated as a toad; and you will be much happier.

Brett W. McCoy wrote:
> Sorry for the off-topicness.  I am looking for a mailing list on
> embedded Perl (like mod_perl is an embedded Perl interpreter in Apache,
>  not like EmbPerl) and can't find anything.  Does anyone know is such a
> list exists?

I don't think such a list exists, but perl-xs could be helpful.
http://lists.perl.org/showlist.cgi?name=perl-xs

See if you can find something else at: http://lists.perl.org/

> Please write me privately so we don't go off-topic on this list.

I'd rather see it answered on the list, so it'll be archived (and possibly
added to the off-site resources page) so in the future people won't need to
ask the same question again.

__________________________________________________________________
Stas Bekman            JAm_pH ------> Just Another mod_perl Hacker
http://stason.org/     mod_perl Guide ---> http://perl.apache.org
mailto:stas@... http://use.perl.org http://apacheweek.com
http://modperlbook.org http://apache.org   http://ticketmaster.com

Stas Bekman wrote:

>> Sorry for the off-topicness.  I am looking for a mailing list on
>> embedded Perl (like mod_perl is an embedded Perl interpreter in
>> Apache,  not like EmbPerl) and can't find anything.  Does anyone know
>> is such a list exists?
>
> I don't think such a list exists, but perl-xs could be helpful.
> http://lists.perl.org/showlist.cgi?name=perl-xs
>
> See if you can find something else at: http://lists.perl.org/

Yeah, the XS list is the closest one I found on that site.  Thanks!

-- Brett
                                            http://www.chapelperilous.net/
------------------------------------------------------------------------
Your boyfriend takes chocolate from strangers.

Sorry for not being clear.

I meant 1.9908 vs. 1.9907. I am using the seen_eos method which is
available from 1.9908.

I have solved this by putting "use modperl 1.9908 ();" in my filter module.
This gives an error upon starting Apache, if the modperl version
is less then required by my module. This is exactly what I wanted to achieve.

Thanks,

Jozef
----- Original Message -----
From: Stas Bekman <stas@...>
Date: Tue, 15 Apr 2003 09:58:56 +1000
To: Jozef Nowak <jnowak@...>
Subject: Re: modperl version check in httpd.conf

> Jozef Nowak wrote:
> > Hello, all.
> >
> > Is it possible to check the version of modperl from the httpd.conf file? I
have the version check in my filter, but would prefer checking the proper
modperl version upon Apache startup.
>
> You mean mp1 vs mp2? or 1.27 vs 1.26?
>
> The former is documented here:
>
http://perl.apache.org/docs/2.0/user/porting/porting.html#Making_Code_Conditiona\
l_on_Running_mod_perl_Version
>
> The latter mechanism doesn't exist, but can be added.
>
> __________________________________________________________________
> Stas Bekman            JAm_pH ------> Just Another mod_perl Hacker
> http://stason.org/     mod_perl Guide ---> http://perl.apache.org
> mailto:stas@... http://use.perl.org http://apacheweek.com
> http://modperlbook.org http://apache.org   http://ticketmaster.com
>

--
_______________________________________________
Sign-up for your own FREE Personalized E-mail at Mail.com
http://www.mail.com/?sr=signup

Hi,

Our company is in a position to take on some Perl contractors (full-time
preferred, part-time possible) in the Atlanta, GA area.
These contracts might potentially turn into full-time employment if
it seems mutually beneficial.

The contract work would be in developing software to extract data from
web sites for an analytical decision support system.  It would be
working for a small company (~ 8 people) which is currently profitable
as an ASP (application service provider) to the travel industry
(car rental and hotel companies).

The work would not be directly related to mod_perl.
It would develop modules which may or may not be run out of a
mod_perl/Apache server.

Please contact me directly (not through the list) if you are interested.

If you are not in the Atlanta metropolitan area and are very interested,
feel free also contact me.  This situation would be less desirable to
us, but it is a possibility.

Please send a resume with any response.

Thanks,

Stephen

Hello, I'm triying to use Apache::DBI 0.91 in an i386 machine running
OpenBSD 3.3 and Apache 1.3.27, I'm using the httpd command
with the -u option to disable the chrooted by default enviroment.

I've loadede my module from the startup.pl script using this line:
use Apache::DBI;

I get only errors, my web server is up, but  this is not running ok,
my httpd process exists but I can not get nothing frommy webserver:

The errors in my  /var/www/logs/error_log are:
[Thu Apr 10 09:34:20 2003] [notice] child pid 10162 exit signal Illegal
instruction (4)
[Thu Apr 10 09:34:21 2003] [notice] child pid 5845 exit signal Illegal
instruction (4)
[Thu Apr 10 09:34:22 2003] [notice] child pid 5293 exit signal Illegal
instruction (4)
[Thu Apr 10 09:34:23 2003] [notice] child pid 28609 exit signal Illegal
instruction (4)
[Thu Apr 10 09:34:24 2003] [notice] child pid 2951 exit signal Illegal
instruction (4)
[Thu Apr 10 09:34:24 2003] [notice] child pid 30535 exit signal Illegal
instruction (4)

After disable Apache::DBI from my startup.pl, I can use my web server
normally, I can use Apache::AuthCookie, Apache::Status, ... .


After this test, I've tried to use Apache::DBI module from the httpd.conf,
I've tried with the next lines:
PerlSetEnv DBI_TRACE "3=/tmp/dbitrace.log"
PerlModule Apache::DBI

I get the same thing:
[Thu Apr 10 09:51:21 2003] [notice] Initializing etag from
/var/www/logs/etag-state
[Thu Apr 10 09:51:21 2003] [notice] Apache/1.3.27 (Unix) mod_perl/1.27
mod_ssl/2.8.12 OpenSSL/0.9.7-beta3 configured -- resuming normal operations
[Thu Apr 10 09:51:21 2003] [notice] Accept mutex: sysvsem (Default: sysvsem)
[Thu Apr 10 09:51:23 2003] [notice] child pid 28014 exit signal Segmentation
fault (11)
[Thu Apr 10 09:51:23 2003] [notice] child pid 7586 exit signal Segmentation
fault (11)
[Thu Apr 10 09:51:23 2003] [notice] child pid 5419 exit signal Segmentation
fault (11)
[Thu Apr 10 09:51:23 2003] [notice] child pid 27794 exit signal Segmentation
fault (11)

I've checked the results in the /tmp/dbitrace.log file, I get the
same message with the 3, 4 and 5 trace level:

    DBI 1.30-nothread dispatch trace level set to 3
    Note: perl is running without the recommended perl -w option
    DBI 1.30-nothread dispatch trace level set to 3
    Note: perl is running without the recommended perl -w option

After that:
I've tried adding the -w option and cleaning a little the
DBI module, but I get the same message without this line:
Note: perl is running without the recommended perl -w option

can some body help me?

thank you.

Have a nice day

--
Alex
alex@...

Eisenhower was very nice,
Nixon was his only vice.
		 -- C. Degen

> > Hello,
> >
> > I'm trying to port a 1.0 authentication module to 2.0 (1.99_08 actually)
> > in Windows using IndigoPerl. With the 1.0 version, a return of REDIRECT to
> > the handler sub would cause a redirection, but that return now does
> > nothing, and the requested file opens directly. I've seen some
> > documentation that mentions REDIRECT is deprecated and to use
> > HTTP_MOVED_TEMPORARILY, but that has the same results. How can I redirect
> > to a different URL using PerlAccessHandler in 2.0?
> >
> > Thanks,
> > Chris
> >
>
>You need to post your code to get help ...
>
>- nick
>
>--

Sorry about that. I'm currently attempting to use the
HTTP_MOVED_TEMPORARILY constant, and the logging messages are all correct
(so  my $redirect_to value is being set correctly), however there is no
redirection at all, it simply opens the requested page. Any ideas are
appreciated.

--------------------------------------------------------------------------------\
-------------------------------------------------

use strict ();
#use Apache::Constants qw(:common REDIRECT FORBIDDEN);
use Apache::Constants qw(:common HTTP_MOVED_TEMPORARILY HTTP_FORBIDDEN);

use Apache::URI ();
use CGI qw(:standard);
use CGI::Cookie ();
use Apache::compat ();


sub handler {
      my $r = shift;

      Apache->request($r);

...

           if (!defined($ourcookie) {
          $r->log_reason("No cookie found (or is unparsable), sending
away.", $r->filename);
          return go_get_a_cookie($r, $redirect_to);
          }
}

sub go_get_a_cookie {

      my ($r, $redirect_to) = @_;

      $r->log_reason("in go_get_a_cookie !\n");

      my $query = $r->parsed_uri->query;
      my $http = $r->dir_config("http_prefix") || (($r->get_server_port ==
443)?"https":"http");
      $redirect_to .= "?$http://" . $r->headers_in->{'Host'} . $r->uri();
      $redirect_to .= "?$query" if ($query);

      $r->header_out(Location => $redirect_to);

      $r->log_reason(" redirect_to = $redirect_to !\n");
      $r->custom_response($Apache::Constants::HTTP_MOVED_TEMPORARILY,"<h1>You
need to Authenticate</h1><p>Please go
<a    href=\"$redirect_to\">$redirect_to</a> to properly authenticate to
this web page.\n");

      return HTTP_MOVED_TEMPORARILY;
}

--------------------------------------------------------------------------------\
--





*nix Administrator/Programmer
Scientific Computing Support
IET, UCDavis
Phone (530)752.1399
Fax     (530)752.0220

On Tuesday, April 15, 2003, at 03:31  AM, Lincoln Stein wrote:
> Hi,
>
> Please find enclosed a proposed release of CGI.pm version 2.92 which
> contains
> fixes from Dan Kogai for UTF-8/perl 5.8, from Stas Bekman for mod_perl
> 2.X
> compatibility, from Steve Hay for broken MSIE browsers, and from Yair
> Lenga
> for non-urlencoded postings.  It also contains a number of fixes from
> yours
> truly (see changelog for details).
>
> If you have a chance, please give this a try on your respective
> systems and
> tell me whether you detect any new anomalies.
>
> Lincoln

Looks good on perl-5.8.0 on MacOS X and FreeBSD-4.8-STABLE.

Dan the Frequent User of CGI.pm

mod_perl doesn't work when run on a file system with acls,
since the code checks manually for a relevant execute bit on
a script before running it (as does the standard mod_cgi
module).  Following is a patch that changes the execute bit
checking to use the posix access() call instead.  A better
solution would probably be to try to execute the file and
return the 'file permissions deny execution' error if
EACCESS is returned.

-hal

[hroberts@h2odev src]$ more mod_perl_acl-1.27.diff
Only in mod_perl_acl-1.27/lib/Apache: #PerlRun.pm#
diff -r -u mod_perl-1.27/lib/Apache/PerlRun.pm
mod_perl_acl-1.27/lib/Apache/PerlRun.pm
--- mod_perl-1.27/lib/Apache/PerlRun.pm Wed May 29 13:14:24 2002
+++ mod_perl_acl-1.27/lib/Apache/PerlRun.pm     Mon Apr  7
18:45:29 2003
@@ -49,7 +49,7 @@
          if (-d _) {
              return DECLINED;
          }
-       unless (-x _ or $Is_Win32) {
+        unless (POSIX::access($r->filename, &POSIX::X_OK)) {
              $r->log_reason("file permissions deny server
execution",
                             $filename);
              return FORBIDDEN;
Only in mod_perl_acl-1.27/lib/Apache: PerlRun.pm~
diff -r -u mod_perl-1.27/lib/Apache/Registry.pm
mod_perl_acl-1.27/lib/Apache/Registry.pm
--- mod_perl-1.27/lib/Apache/Registry.pm        Thu May 23
00:00:44 2002
+++ mod_perl_acl-1.27/lib/Apache/Registry.pm    Mon Apr  7
18:45:37 2003
@@ -56,7 +56,7 @@
                             $filename);
              return FORBIDDEN;
          }
-       unless (-x _ or $Is_Win32) {
+        unless (POSIX::access($r->filename, &POSIX::X_OK)) {
              $r->log_reason("file permissions deny server
execution",
                             $filename);
              return FORBIDDEN;
Only in mod_perl_acl-1.27/lib/Apache: Registry.pm~
diff -r -u mod_perl-1.27/src/modules/perl/PerlRunXS.xs
mod_perl_acl-1.27/src/modules/perl/PerlRunXS.xs
--- mod_perl-1.27/src/modules/perl/PerlRunXS.xs Thu Mar  2
03:30:40 2000
+++ mod_perl_acl-1.27/src/modules/perl/PerlRunXS.xs     Mon
Apr  7 17:57:57 2003
@@ -177,10 +177,10 @@
       if (S_ISDIR(r->finfo.st_mode)) {
          return DECLINED;
       }
-    if (!can_exec(&r->finfo)) {
-       log_scripterror(r, FORBIDDEN,
-                       "file permissions deny server
execution");
-    }
+    if (access(r->filename, X_OK)) {
+      return log_scripterror(r, conf, FORBIDDEN,
APLOG_NOERRNO,
+                             "file permissions deny server
execution");
+    }
       return OK;
   }

> Rather than guessing what could be the problem, why not get a core
> file and look at its trace?

After a bit of playing with the system settings, we were able to get
some core files. We had to use the coreadm utility on Solaris to allow
a setuid process to leave a core file.

Anyway, here are the results of one of the core dumps from gdb:

#0  0x00139f38 in Perl_op_const_sv ()
#1  0x0016b388 in Perl_sv_setsv ()
#2  0x0015e1fc in Perl_pp_sassign ()
#3  0x0015dc80 in Perl_runops_standard ()
#4  0x0010dfc8 in S_call_body ()
#5  0x0010dcc8 in perl_call_sv ()
#6  0x0010d7e0 in perl_call_method ()
#7  0x0003204c in perl_call_handler (sv=0x36cee4, r=0x392040, args=0x0)
at mod_perl.c:1655
#8  0x000310a0 in perl_run_stacked_handlers (hook=0x1b55b8
"PerlHandler", r=0x392040, handlers=0x36cf50) at mod_perl.c:1371
#9  0x0002d620 in perl_handler (r=0x392040) at mod_perl.c:897
#10 0x000a787c in ap_invoke_handler (r=0x392040) at http_config.c:518
#11 0x000c6970 in process_request_internal (r=0x392040) at
http_request.c:1308
#12 0x000c69f4 in ap_process_request (r=0x392040) at http_request.c:1324
#13 0x000b95b4 in child_main (child_num_arg=3) at http_main.c:4603
#14 0x000b9940 in make_child (s=0x201840, slot=3, now=1050426606) at
http_main.c:4773
#15 0x000b9a64 in startup_children (number_to_start=12) at
http_main.c:4800
#16 0x000ba494 in standalone_main (argc=1, argv=0xffbefb04) at
http_main.c:5108
#17 0x000bb0f0 in main (argc=1, argv=0xffbefb04) at http_main.c:5456


==========================================
Jim Brandt
Administrative Computing Services
University at Buffalo

Hello,

I'm re-installing modules after an upgrade of my ibook from 10.1.5 to
10.2.5 and get the error shown below when "make test" runs for
Apache::DBI.  This is the version of perl and mod_perl that comes with
jaguar - perl v5.6.0, Apache/1.3.27 (Darwin) mod_perl/1.26.  I run cgi
scripts through mod_perl - mod_perl seems to be working OK.  I noticed
that during the DBI install - it kept complaining about the 5.6.0
version of perl.

Any suggestions would be appreciated.


Thanks,

Warren


Here's the test(s) - the make test output follows

[t]$ cat 10mysql.t
use strict;
use Test::More tests => 7;

BEGIN {
    # trick DBI.pm into thinking we are running under mod_perl
    $ENV{GATEWAY_INTERFACE} = 'CGI-Perl';
    use_ok('Apache::DBI', 'load Apache::DBI')
};

my $dbd_mysql = eval { require DBD::mysql };

#$Apache::DBI::DEBUG = 10;

SKIP: {
    skip "Could not load DBD::mysql", 6 unless $dbd_mysql;

    ok($dbd_mysql, "DBD::mysql loaded");

    my $dbh_1 = DBI->connect('dbi:mysql:test:localhost', 'root',
'b1gh3z', { RaiseError => 0, PrintError => 0 });

   SKIP: {
      skip "Could not connect to test database: $DBI::errstr", 5 unless
$dbh_1;
      ok(my $thread_1 = $dbh_1->{'mysql_thread_id'}, "Connected 1");

      my $dbh_2 = DBI->connect('dbi:mysql:test:localhost', 'root',
'b1gh3z', { RaiseError => 0, PrintError => 0 });
      ok(my $thread_2 = $dbh_2->{'mysql_thread_id'}, "Connected 2");

      is($thread_1, $thread_2, "got the same connection both times");

      my $dbh_3 = DBI->connect('dbi:mysql:test:localhost', 'root',
'b1gh3z', { RaiseError => 0, PrintError => 1 });
      ok(my $thread_3 = $dbh_3->{'mysql_thread_id'}, "Connected 3");

      isnt($thread_1, $thread_3, "got different connection from different
attributes");


    }


}

1;

+++++++++++++++++++++++++++++++++++

[Apache-DBI-0.91]$ sudo make test
PERL_DL_NONLAZY=1 /usr/bin/perl -Iblib/arch -Iblib/lib
-I/System/Library/Perl/darwin -I/System/Library/Perl -e 'use
Test::Harness qw(&runtests $verbose); $verbose=0; runtests @ARGV;' t/*.t
t/10mysql....NOK 5#     Failed test (t/10mysql.t at line 28)
#          got: '14'
#     expected: '15'
t/10mysql....ok 7/7# Looks like you failed 1 tests of 7.
t/10mysql....dubious
          Test returned status 1 (wstat 256, 0x100)
DIED. FAILED test 5
          Failed 1/7 tests, 85.71% okay
Failed Test Stat Wstat Total Fail  Failed  List of Failed
------------------------------------------------------------------------
-------
t/10mysql.t    1   256     7    1  14.29%  5
Failed 1/1 test scripts, 0.00% okay. 1/7 subtests failed, 85.71% okay.
make: *** [test_dynamic] Error 2

++++++++++++++++++++++++++++++++++++

Warren Pollans wrote:
> Hello,
>
> I'm re-installing modules after an upgrade of my ibook from 10.1.5 to
> 10.2.5 and get the error shown below when "make test" runs for
> Apache::DBI.  This is the version of perl and mod_perl that comes with
> jaguar - perl v5.6.0, Apache/1.3.27 (Darwin) mod_perl/1.26.  I run cgi
> scripts through mod_perl - mod_perl seems to be working OK.  I noticed
> that during the DBI install - it kept complaining about the 5.6.0
> version of perl.

It complains for a reason. perl-5.6.0 is *very* buggy, so if you can update to
at least 5.6.1 chances are that all your problems will go away. Care to try?

__________________________________________________________________
Stas Bekman            JAm_pH ------> Just Another mod_perl Hacker
http://stason.org/     mod_perl Guide ---> http://perl.apache.org
mailto:stas@... http://use.perl.org http://apacheweek.com
http://modperlbook.org http://apache.org   http://ticketmaster.com

Jozef Nowak wrote:
> Sorry for not being clear.
>
> I meant 1.9908 vs. 1.9907. I am using the seen_eos method which is
> available from 1.9908.
>
> I have solved this by putting "use modperl 1.9908 ();" in my filter module.
> This gives an error upon starting Apache, if the modperl version
> is less then required by my module. This is exactly what I wanted to achieve.

Cool, it was documented here:
http://perl.apache.org/docs/2.0/user/porting/porting.html#Requiring_a_specific_m\
od_perl_version_
Now I'll add the trick to require a specific beta sub-version. Thanks!

__________________________________________________________________
Stas Bekman            JAm_pH ------> Just Another mod_perl Hacker
http://stason.org/     mod_perl Guide ---> http://perl.apache.org
mailto:stas@... http://use.perl.org http://apacheweek.com
http://modperlbook.org http://apache.org   http://ticketmaster.com

Jim Brandt wrote:
>> Rather than guessing what could be the problem, why not get a core
>> file and look at its trace?
>
>
> After a bit of playing with the system settings, we were able to get
> some core files. We had to use the coreadm utility on Solaris to allow a
> setuid process to leave a core file.

was this process any different from the one documented here?
http://perl.apache.org/docs/2.0/devel/debug/c.html#Obtaining_core_Files_under_So\
laris

> Anyway, here are the results of one of the core dumps from gdb:
>
> #0  0x00139f38 in Perl_op_const_sv ()
> #1  0x0016b388 in Perl_sv_setsv ()
> #2  0x0015e1fc in Perl_pp_sassign ()
> #3  0x0015dc80 in Perl_runops_standard ()
> #4  0x0010dfc8 in S_call_body ()
> #5  0x0010dcc8 in perl_call_sv ()
> #6  0x0010d7e0 in perl_call_method ()

So you can see that the problem happens on the perl side. However since your
perl was compiled without -Doptimize='-g' the trace doesn't show the
arguments, so it's hard to tell what was the cause.

> #7  0x0003204c in perl_call_handler (sv=0x36cee4, r=0x392040, args=0x0)
> at mod_perl.c:1655
> #8  0x000310a0 in perl_run_stacked_handlers (hook=0x1b55b8
> "PerlHandler", r=0x392040, handlers=0x36cf50) at mod_perl.c:1371
> #9  0x0002d620 in perl_handler (r=0x392040) at mod_perl.c:897
> #10 0x000a787c in ap_invoke_handler (r=0x392040) at http_config.c:518
> #11 0x000c6970 in process_request_internal (r=0x392040) at
> http_request.c:1308
> #12 0x000c69f4 in ap_process_request (r=0x392040) at http_request.c:1324
> #13 0x000b95b4 in child_main (child_num_arg=3) at http_main.c:4603
> #14 0x000b9940 in make_child (s=0x201840, slot=3, now=1050426606) at
> http_main.c:4773
> #15 0x000b9a64 in startup_children (number_to_start=12) at http_main.c:4800
> #16 0x000ba494 in standalone_main (argc=1, argv=0xffbefb04) at
> http_main.c:5108
> #17 0x000bb0f0 in main (argc=1, argv=0xffbefb04) at http_main.c:5456

__________________________________________________________________
Stas Bekman            JAm_pH ------> Just Another mod_perl Hacker
http://stason.org/     mod_perl Guide ---> http://perl.apache.org
mailto:stas@... http://use.perl.org http://apacheweek.com
http://modperlbook.org http://apache.org   http://ticketmaster.com

Hal Roberts wrote:
> mod_perl doesn't work when run on a file system with acls, since the
> code checks manually for a relevant execute bit on a script before
> running it (as does the standard mod_cgi module).  Following is a patch
> that changes the execute bit checking to use the posix access() call
> instead.

That's pretty cool, Hal! Thanks.

> A better solution would probably be to try to execute the file
> and return the 'file permissions deny execution' error if EACCESS is
> returned.

That would be pretty dangerous, how can you try without actually executing the
file? It'd also add a big overhead to the run time.

A few comments on the patch (mainly worried about this not working on all
platforms mod_perl is running on).

> [hroberts@h2odev src]$ more mod_perl_acl-1.27.diff
> Only in mod_perl_acl-1.27/lib/Apache: #PerlRun.pm#
> diff -r -u mod_perl-1.27/lib/Apache/PerlRun.pm
> mod_perl_acl-1.27/lib/Apache/PerlRun.pm
> --- mod_perl-1.27/lib/Apache/PerlRun.pm Wed May 29 13:14:24 2002
> +++ mod_perl_acl-1.27/lib/Apache/PerlRun.pm     Mon Apr  7 18:45:29 2003
> @@ -49,7 +49,7 @@
>         if (-d _) {
>             return DECLINED;
>         }
> -       unless (-x _ or $Is_Win32) {

require POSIX; is missing.

> +        unless (POSIX::access($r->filename, &POSIX::X_OK)) {
>             $r->log_reason("file permissions deny server execution",
>                            $filename);
>             return FORBIDDEN;
> Only in mod_perl_acl-1.27/lib/Apache: PerlRun.pm~
> diff -r -u mod_perl-1.27/lib/Apache/Registry.pm
> mod_perl_acl-1.27/lib/Apache/Registry.pm
> --- mod_perl-1.27/lib/Apache/Registry.pm        Thu May 23 00:00:44 2002
> +++ mod_perl_acl-1.27/lib/Apache/Registry.pm    Mon Apr  7 18:45:37 2003
> @@ -56,7 +56,7 @@
>                            $filename);
>             return FORBIDDEN;
>         }
> -       unless (-x _ or $Is_Win32) {
> +        unless (POSIX::access($r->filename, &POSIX::X_OK)) {
>             $r->log_reason("file permissions deny server execution",
>                            $filename);
>             return FORBIDDEN;
> Only in mod_perl_acl-1.27/lib/Apache: Registry.pm~
> diff -r -u mod_perl-1.27/src/modules/perl/PerlRunXS.xs
> mod_perl_acl-1.27/src/modules/perl/PerlRunXS.xs
> --- mod_perl-1.27/src/modules/perl/PerlRunXS.xs Thu Mar  2 03:30:40 2000
> +++ mod_perl_acl-1.27/src/modules/perl/PerlRunXS.xs     Mon Apr  7
> 17:57:57 2003
> @@ -177,10 +177,10 @@
>      if (S_ISDIR(r->finfo.st_mode)) {
>         return DECLINED;
>      }
> -    if (!can_exec(&r->finfo)) {
> -       log_scripterror(r, FORBIDDEN,
> -                       "file permissions deny server execution");
> -    }
> +    if (access(r->filename, X_OK)) {

need to make sure that we have:

       #include <unistd.h>

(that's what access(2) manpage on linux says to use)

is this the same header file on all platforms supporting POSIX? POSIX.xs has:

#ifdef I_UNISTD
#include <unistd.h>
#endif

so it's probably not so simple.

> +      return log_scripterror(r, conf, FORBIDDEN, APLOG_NOERRNO,
> +                             "file permissions deny server execution");
> +    }
>      return OK;
>  }

Finally, the perl POSIX manpage says:

         Furthermore, some evil vendors will claim 1003.1 compliance, but in
         fact are not so: they will not pass the PCTS (POSIX Compliance Test
         Suites).  For example, one vendor may not define EDEADLK, or the seman-
         tics of the errno values set by open(2) might not be quite right.  Perl
         does not attempt to verify POSIX compliance.  That means you can cur-
         rently successfully say "use POSIX",  and then later in your program
         you find that your vendor has been lax and there's no usable ICANON
         macro after all.  This could be construed to be a bug.

Any potential problems here? I haven't used the access() interface before, so
if someone can confirm that it's cross-platform it'd help.

__________________________________________________________________
Stas Bekman            JAm_pH ------> Just Another mod_perl Hacker
http://stason.org/     mod_perl Guide ---> http://perl.apache.org
mailto:stas@... http://use.perl.org http://apacheweek.com
http://modperlbook.org http://apache.org   http://ticketmaster.com

Lincoln Stein wrote:
> I've implemented the ability to pass $r to CGI and CGI::Cookie on
> initialization.  This should make it easier to use CGI in earlier phases of
> the request cycle.  It works for me, but please try it.

I've just read the diff, hadn't tried it yet. It looks good, but there are
several things that need more work. In mod_perl 2, ref($r) returns
'Apache::RequestRec' so it should be:

... && ($r->isa('Apache') || $r->isa('Apache::RequestRec'))

in both modules. or have the logic different depending on the value of
$MOD_PERL.

The addition of the r() API is probably worthy an entry in the Changes log.

Also why using Apache->request($r) for the functional API and not CGI::r()? We
should stay way from using Apache->request if possible, because this requires
storing/retrieving data from the thread-local storage in the threaded
environment. The functional interface is already maintaining a set of global
vars, so there shouldn't be a problem adding a new global.

Finally CGI::Cookie uses $r->header_in, which is deprecated in mp1 and doesn't
exists in mp2. It should be $r->headers_in->get('Cookie'), or at least do that
for mp2.

Nick, it looks like a good time to plug your CGI::Cookie patch in.

Since now we make sure that we have $r for both mod_perl generations, the next
stage for CGI.pm to stop relying on %ENV and use $r->headers_in interface to
get the required request header values.

__________________________________________________________________
Stas Bekman            JAm_pH ------> Just Another mod_perl Hacker
http://stason.org/     mod_perl Guide ---> http://perl.apache.org
mailto:stas@... http://use.perl.org http://apacheweek.com
http://modperlbook.org http://apache.org   http://ticketmaster.com

Hi,

Is it worth outputting the log details for web sites out to the database
using mod perl, or is it best to use the general logging to a file.

Also would it be possible to log to a different DB for each site, or even
for each part of a site.

Thanks,

Adam



---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.471 / Virus Database: 269 - Release Date: 13/04/2003

Hi,

Which is the best template system to use with mod perl. I have looked
through the information on the mod perl site, and there a quite a few
listed, it also appears that most of the template output is cached, which
could cause us a problem.

We are looking at using a template system that will connect to a DB and
locate the correct template files to use dependent on a query and then use
those files for the output. Would this be possible?

Thanks,

Adam


---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.471 / Virus Database: 269 - Release Date: 13/04/2003

Hi there,

On Wed, 16 Apr 2003, Adam Gent wrote:

> Is it worth outputting the log details for web sites out to the database
> using mod perl, or is it best to use the general logging to a file.

AFAICT, at the moment you are the only person who has the information
needed to answer that question, which will obviously depend upon the
particular situation.  If you can give more information you might get
a more helpful answer.

> Also would it be possible to log to a different DB for each site, or even
> for each part of a site.

Yes.  But answer the first question first, then trawl around in the
archives for more information, and for example on CPAN for modules to
log to a database.  You can also check out the Apache documentation.
Apache has a flexible logging system.

73,
Ged.

Hello again,

On Wed, 16 Apr 2003, Adam Gent wrote:

> Which is the best template system to use with mod perl.

Which is the best margarine to use with rye bread?

Please do at least some homework before you ask impossible questions.
This particular topic comes up quite a lot on this List, if you check
the archives that should keep you in bed-time reading for weeks.

73,
Ged.

Once you've decided on your templating system, you will then need to work
out how best to use that system.
Look at the pages you are trying to make, look at how much redundancy there
is across pages.

But like Ged said, read,read,read!

Marty
----- Original Message -----
From: "Ged Haywood" <ged@...>
To: "Adam Gent" <apache-mail@...>
Cc: <modperl@...>
Sent: Wednesday, April 16, 2003 1:15 PM
Subject: Re: Template Systems


> Hello again,
>
> On Wed, 16 Apr 2003, Adam Gent wrote:
>
> > Which is the best template system to use with mod perl.
>
> Which is the best margarine to use with rye bread?
>
> Please do at least some homework before you ask impossible questions.
> This particular topic comes up quite a lot on this List, if you check
> the archives that should keep you in bed-time reading for weeks.
>
> 73,
> Ged.
>
>

Adam Gent wrote:
> Which is the best template system to use with mod perl.

If you haven't read this yet, read it now:
http://perl.apache.org/docs/tutorials/tmpl/comparison/comparison.html

If you still have no idea, explain your requirements a bit so we can
narrow it down.

> I have looked
> through the information on the mod perl site, and there a quite a few
> listed, it also appears that most of the template output is cached, which
> could cause us a problem.

No, most of the templates are cached.  Output caching is a somewhat
unusual feature, and no system will cache output unless you tell it to.

> We are looking at using a template system that will connect to a DB and
> locate the correct template files to use dependent on a query and then use
> those files for the output. Would this be possible?

Yes, almost any system can do that.  You just have to do the database
work and find the template name, and then you can tell the templating
system to run that template.

- Perrin

On Wed, 16 Apr 2003, Perrin Harkins wrote:

> > I have looked
> > through the information on the mod perl site, and there a quite a few
> > listed, it also appears that most of the template output is cached, which
> > could cause us a problem.
>
> No, most of the templates are cached.  Output caching is a somewhat
> unusual feature, and no system will cache output unless you tell it to.

Except AxKit :-)

--
<!-- Matt -->
<:->get a SMart net</:->
Spam trap - do not mail: spam-sig@...

On Thu, 10 Apr 2003, Chris Pelton wrote:

>
>
> > > Hello,
> > >
> > > I'm trying to port a 1.0 authentication module to 2.0 (1.99_08 actually)
> > > in Windows using IndigoPerl. With the 1.0 version, a return of REDIRECT to
> > > the handler sub would cause a redirection, but that return now does
> > > nothing, and the requested file opens directly. I've seen some
> > > documentation that mentions REDIRECT is deprecated and to use
> > > HTTP_MOVED_TEMPORARILY, but that has the same results. How can I redirect
> > > to a different URL using PerlAccessHandler in 2.0?
> > >
> > > Thanks,
> > > Chris
> > >
> >
> >You need to post your code to get help ...
> >
> >- nick
> >
> >--
>
> Sorry about that. I'm currently attempting to use the
> HTTP_MOVED_TEMPORARILY constant, and the logging messages are all correct
> (so  my $redirect_to value is being set correctly), however there is no
> redirection at all, it simply opens the requested page. Any ideas are
> appreciated.
>
>
--------------------------------------------------------------------------------\
-------------------------------------------------
>
> use strict ();
> #use Apache::Constants qw(:common REDIRECT FORBIDDEN);
> use Apache::Constants qw(:common HTTP_MOVED_TEMPORARILY HTTP_FORBIDDEN);
>
> use Apache::URI ();
> use CGI qw(:standard);
> use CGI::Cookie ();
> use Apache::compat ();
>
>
> sub handler {
>      my $r = shift;
>
>      Apache->request($r);
>
> ...
>
>           if (!defined($ourcookie) {
>          $r->log_reason("No cookie found (or is unparsable), sending
> away.", $r->filename);
>          return go_get_a_cookie($r, $redirect_to);
>          }
> }
>
> sub go_get_a_cookie {
>
>      my ($r, $redirect_to) = @_;
>
>      $r->log_reason("in go_get_a_cookie !\n");
>
>      my $query = $r->parsed_uri->query;
>      my $http = $r->dir_config("http_prefix") || (($r->get_server_port ==
> 443)?"https":"http");
>      $redirect_to .= "?$http://" . $r->headers_in->{'Host'} . $r->uri();
>      $redirect_to .= "?$query" if ($query);
>
>      $r->header_out(Location => $redirect_to);
>
>      $r->log_reason(" redirect_to = $redirect_to !\n");
>      $r->custom_response($Apache::Constants::HTTP_MOVED_TEMPORARILY,"<h1>You
> need to Authenticate</h1><p>Please go
> <a    href=\"$redirect_to\">$redirect_to</a> to properly authenticate to
> this web page.\n");
>
>      return HTTP_MOVED_TEMPORARILY;

Try:
return Apache::return HTTP_MOVED_TEMPORARILY;

That's what's required in mp2.

You are using Apache::compat so I am not sure ... You should be able to run
mp1 code unchanged, but since I do not use it I cannot vouch for it.

I would port your module to native mp2 code if I were you. Just take out the
call to Apache::compat and use MethodLookup as documented to chnage the
necessary
statements and calls, and you'll have it done in a flash.


> }
>
>
--------------------------------------------------------------------------------\
--
>
>
>
>
>
> *nix Administrator/Programmer
> Scientific Computing Support
> IET, UCDavis
> Phone (530)752.1399
> Fax     (530)752.0220
>
>


- nick

--

~~~~~~~~~~~~~~~~~~~~
Nick Tonkin   {|8^)>

> was this process any different from the one documented here?
> http://perl.apache.org/docs/2.0/devel/debug/
> c.html#Obtaining_core_Files_under_Solaris
>

I think a blurb like this would help:

By default, Solaris 8 won't allow a setuid process to write a core file
to the file system. Since apache starts as root and spawns children as
'nobody', core dumps won't produce core files unless you modify the
system settings.

To see the current settings, run the coreadm command with no parameters
and you'll see:

%coreadm
       global core file pattern:
         init core file pattern: core
              global core dumps: disabled
         per-process core dumps: enabled
        global setid core dumps: disabled
   per-process setid core dumps: disabled
       global core dump logging: disabled

These settings are stored in the /etc/coreadm.conf file, but you should
set them with the coreadm utility. As super-user, you can run coreadm
with -g to set the pattern and path for core files (you can use a few
variables here) and -e to enable some of the disabled items. After
setting a new pattern, enabling global, global-setid, and log, and
rebooting the system (reboot is required), the new settings look like:

%coreadm
       global core file pattern: /usr/local/apache/cores/core.%f.%p
         init core file pattern: core
              global core dumps: enabled
         per-process core dumps: enabled
        global setid core dumps: enabled
   per-process setid core dumps: disabled
       global core dump logging: enabled

Now you'll start to see core files in the designated cores directory
and they will look like core.httpd.2222 where httpd is the name of the
executable and the 2222 is the process id. The new core files will be
read/write for root only to maintain some security, and you should
probably do this on development systems only.

==========================================
Jim Brandt
Administrative Computing Services
University at Buffalo

On Wed, 16 Apr 2003, Stas Bekman wrote:

> Lincoln Stein wrote:
> > I've implemented the ability to pass $r to CGI and CGI::Cookie on
> > initialization.  This should make it easier to use CGI in earlier phases of
> > the request cycle.  It works for me, but please try it.
>
> I've just read the diff, hadn't tried it yet. It looks good, but there are
> several things that need more work. In mod_perl 2, ref($r) returns
> 'Apache::RequestRec' so it should be:
>
> ... && ($r->isa('Apache') || $r->isa('Apache::RequestRec'))
>
> in both modules. or have the logic different depending on the value of
$MOD_PERL.
>
> The addition of the r() API is probably worthy an entry in the Changes log.
>
> Also why using Apache->request($r) for the functional API and not CGI::r()? We
> should stay way from using Apache->request if possible, because this requires
> storing/retrieving data from the thread-local storage in the threaded
> environment. The functional interface is already maintaining a set of global
> vars, so there shouldn't be a problem adding a new global.
>
> Finally CGI::Cookie uses $r->header_in, which is deprecated in mp1 and doesn't
> exists in mp2. It should be $r->headers_in->get('Cookie'), or at least do that
> for mp2.
>
> Nick, it looks like a good time to plug your CGI::Cookie patch in.

Alright. Where should I send it? Lincoln, do you want it?

>
> Since now we make sure that we have $r for both mod_perl generations, the next
> stage for CGI.pm to stop relying on %ENV and use $r->headers_in interface to
> get the required request header values.
>
> __________________________________________________________________
> Stas Bekman            JAm_pH ------> Just Another mod_perl Hacker
> http://stason.org/     mod_perl Guide ---> http://perl.apache.org
> mailto:stas@... http://use.perl.org http://apacheweek.com
> http://modperlbook.org http://apache.org   http://ticketmaster.com
>


- nick

--

~~~~~~~~~~~~~~~~~~~~
Nick Tonkin   {|8^)>