2.0.8 POST problem

RT Users
1

Hello–

I’ve been happily using RT1 for almost three years. Today, I decided to
make the jump to RT2. The installation went well, and everything from my
RT1 instance imported perfectly. But there is some problem with any RT2
form that uses the POST method instead of GET. With the main login screen,
for instance, if I try to use the username and password form elements, I
just get dumped back to the same screen. However, if I tag on
?user=xxxx&pass=yyyy to the base RT2 URL, I get in just fine. It seems
that most of the other forms use GET, so the system is functional, but
several of the administrative forms use POST.

I have checked the list archives. There have been a few threads talking
about this problem, but no solutions aside from changing all of the POSTs
in the code over to GETs, which seems clunky.

If I look in my webserver logs, it looks like the browser is actually
sending two separate requests, almost as if the HTTP headers were getting
messed up:

169.237.xxx.yyy - - [18/Oct/2001:13:39:01 -0700] “POST /rt2/ HTTP/1.1” 200
2071
169.237.xxx.yyy - - [18/Oct/2001:13:39:01 -0700] “user=test&pass=testGET
/rt2/NoAuth/webrt.css HTTP/1.1” 200 1484

This problem is consistent across several different browsers, including
Mozilla, IE 5.5, and Netscape 4.7.

I’m using RT 2.0.8, Apache 1.3.19, Perl 5.6.1, mod_perl 1.26 and fresh
builds of all the required Perl modules (compiled from CPAN today) all on
a RedHat 7.0 system.

Has anybody else seen this behaviour (and, hopefully, a solution)?

Thanks,

Jeremy.
Jeremy M. Smith jeremy@ucdavis.edu
Technical Support Manager, Programmer
University of California, Davis
Languages and Literatures
Language Learning Center

2

“Jeremy M. Smith” wrote:

Hello–

I’ve been happily using RT1 for almost three years. Today, I decided to
make the jump to RT2. The installation went well, and everything from my
RT1 instance imported perfectly. But there is some problem with any RT2
form that uses the POST method instead of GET. With the main login screen,
for instance, if I try to use the username and password form elements, I
just get dumped back to the same screen. However, if I tag on
?user=xxxx&pass=yyyy to the base RT2 URL, I get in just fine. It seems
that most of the other forms use GET, so the system is functional, but
several of the administrative forms use POST.

I have checked the list archives. There have been a few threads talking
about this problem, but no solutions aside from changing all of the POSTs
in the code over to GETs, which seems clunky.

If I look in my webserver logs, it looks like the browser is actually
sending two separate requests, almost as if the HTTP headers were getting
messed up:

169.237.xxx.yyy - - [18/Oct/2001:13:39:01 -0700] “POST /rt2/ HTTP/1.1” 200
2071
169.237.xxx.yyy - - [18/Oct/2001:13:39:01 -0700] “user=test&pass=testGET
/rt2/NoAuth/webrt.css HTTP/1.1” 200 1484

This problem is consistent across several different browsers, including
Mozilla, IE 5.5, and Netscape 4.7.

I’m using RT 2.0.8, Apache 1.3.19, Perl 5.6.1, mod_perl 1.26 and fresh
builds of all the required Perl modules (compiled from CPAN today) all on
a RedHat 7.0 system.

Has anybody else seen this behaviour (and, hopefully, a solution)?

Thats two of us that have the same problem - but the only difference is
that I am using Apache 1.3.20, and RedHat 7.1.

The only kludge I have is to change the posts to gets.

Any else any ideas ?

Greg

3

Are you guys using redhat’s builds of mod_perl and apache?On Fri, Oct 19, 2001 at 08:54:11AM +0000, Greg Cope wrote:

“Jeremy M. Smith” wrote:

Hello–

I’ve been happily using RT1 for almost three years. Today, I decided to
make the jump to RT2. The installation went well, and everything from my
RT1 instance imported perfectly. But there is some problem with any RT2
form that uses the POST method instead of GET. With the main login screen,
for instance, if I try to use the username and password form elements, I
just get dumped back to the same screen. However, if I tag on
?user=xxxx&pass=yyyy to the base RT2 URL, I get in just fine. It seems
that most of the other forms use GET, so the system is functional, but
several of the administrative forms use POST.

I have checked the list archives. There have been a few threads talking
about this problem, but no solutions aside from changing all of the POSTs
in the code over to GETs, which seems clunky.

If I look in my webserver logs, it looks like the browser is actually
sending two separate requests, almost as if the HTTP headers were getting
messed up:

169.237.xxx.yyy - - [18/Oct/2001:13:39:01 -0700] “POST /rt2/ HTTP/1.1” 200
2071
169.237.xxx.yyy - - [18/Oct/2001:13:39:01 -0700] “user=test&pass=testGET
/rt2/NoAuth/webrt.css HTTP/1.1” 200 1484

This problem is consistent across several different browsers, including
Mozilla, IE 5.5, and Netscape 4.7.

I’m using RT 2.0.8, Apache 1.3.19, Perl 5.6.1, mod_perl 1.26 and fresh
builds of all the required Perl modules (compiled from CPAN today) all on
a RedHat 7.0 system.

Has anybody else seen this behaviour (and, hopefully, a solution)?

Thats two of us that have the same problem - but the only difference is
that I am using Apache 1.3.20, and RedHat 7.1.

The only kludge I have is to change the posts to gets.

Any else any ideas ?

Greg

rt-users mailing list
rt-users@lists.fsck.com
http://lists.fsck.com/mailman/listinfo/rt-users

http://www.bestpractical.com/products/rt – Trouble Ticketing. Free.

4

Nope. I’m using my own build of both. The most suspect thing in my setup
is that I’m building mod_perl as a DSO in order to preserve some other
complex apache mods.

Jeremy.

Are you guys using redhat’s builds of mod_perl and apache?

“Jeremy M. Smith” wrote:

Hello–
I’ve been happily using RT1 for almost three years. Today, I
decided to
make the jump to RT2. The installation went well, and everything from
my
RT1 instance imported perfectly. But there is some problem with any
RT2
form that uses the POST method instead of GET. With the main login
screen,
for instance, if I try to use the username and password form elements,
I
just get dumped back to the same screen. However, if I tag on
?user=xxxx&pass=yyyy to the base RT2 URL, I get in just fine. It seems
that most of the other forms use GET, so the system is functional, but
several of the administrative forms use POST.
I have checked the list archives. There have been a few threads
talking
about this problem, but no solutions aside from changing all of the
POSTs
in the code over to GETs, which seems clunky.
If I look in my webserver logs, it looks like the browser is
actually
sending two separate requests, almost as if the HTTP headers were
getting
messed up:
169.237.xxx.yyy - - [18/Oct/2001:13:39:01 -0700] “POST /rt2/
HTTP/1.1” 200
2071
169.237.xxx.yyy - - [18/Oct/2001:13:39:01 -0700]
“user=test&pass=testGET
/rt2/NoAuth/webrt.css HTTP/1.1” 200 1484
This problem is consistent across several different browsers,
including
Mozilla, IE 5.5, and Netscape 4.7.
I’m using RT 2.0.8, Apache 1.3.19, Perl 5.6.1, mod_perl 1.26 and
fresh
builds of all the required Perl modules (compiled from CPAN today) all
on
a RedHat 7.0 system.
Has anybody else seen this behaviour (and, hopefully, a solution)?
Thats two of us that have the same problem - but the only
difference is
that I am using Apache 1.3.20, and RedHat 7.1.
The only kludge I have is to change the posts to gets.
Any else any ideas ?
Greg

rt-users mailing list
rt-users@lists.fsck.com
http://lists.fsck.com/mailman/listinfo/rt-users

http://www.bestpractical.com/products/rt – Trouble Ticketing. Free

Jeremy M. Smith jeremy@ucdavis.edu
Technical Support Manager, Programmer
University of California, Davis
Languages and Literatures
Language Learning Center

5

Rather. mod_perl as a DSO is not going to be any fun. If you need mod_perl
as a DSO, you should probably use the fastcgi handler for the RT web interface.

JesseOn Fri, Oct 19, 2001 at 08:15:49AM -0700, Jeremy M. Smith wrote:

Nope. I’m using my own build of both. The most suspect thing in my setup
is that I’m building mod_perl as a DSO in order to preserve some other
complex apache mods.

Jeremy.

Are you guys using redhat’s builds of mod_perl and apache?

On Fri, Oct 19, 2001 at 08:54:11AM +0000, Greg Cope wrote:

“Jeremy M. Smith” wrote:

Hello–
I’ve been happily using RT1 for almost three years. Today, I
decided to
make the jump to RT2. The installation went well, and everything from
my
RT1 instance imported perfectly. But there is some problem with any
RT2
form that uses the POST method instead of GET. With the main login
screen,
for instance, if I try to use the username and password form elements,
I
just get dumped back to the same screen. However, if I tag on
?user=xxxx&pass=yyyy to the base RT2 URL, I get in just fine. It seems
that most of the other forms use GET, so the system is functional, but
several of the administrative forms use POST.
I have checked the list archives. There have been a few threads
talking
about this problem, but no solutions aside from changing all of the
POSTs
in the code over to GETs, which seems clunky.
If I look in my webserver logs, it looks like the browser is
actually
sending two separate requests, almost as if the HTTP headers were
getting
messed up:
169.237.xxx.yyy - - [18/Oct/2001:13:39:01 -0700] “POST /rt2/
HTTP/1.1” 200
2071
169.237.xxx.yyy - - [18/Oct/2001:13:39:01 -0700]
“user=test&pass=testGET
/rt2/NoAuth/webrt.css HTTP/1.1” 200 1484
This problem is consistent across several different browsers,
including
Mozilla, IE 5.5, and Netscape 4.7.
I’m using RT 2.0.8, Apache 1.3.19, Perl 5.6.1, mod_perl 1.26 and
fresh
builds of all the required Perl modules (compiled from CPAN today) all
on
a RedHat 7.0 system.
Has anybody else seen this behaviour (and, hopefully, a solution)?
Thats two of us that have the same problem - but the only
difference is
that I am using Apache 1.3.20, and RedHat 7.1.
The only kludge I have is to change the posts to gets.
Any else any ideas ?
Greg

rt-users mailing list
rt-users@lists.fsck.com
http://lists.fsck.com/mailman/listinfo/rt-users

–
http://www.bestpractical.com/products/rt – Trouble Ticketing. Free

–

Jeremy M. Smith jeremy@ucdavis.edu
Technical Support Manager, Programmer
University of California, Davis
Languages and Literatures
Language Learning Center

rt-users mailing list
rt-users@lists.fsck.com
http://lists.fsck.com/mailman/listinfo/rt-users

http://www.bestpractical.com/products/rt – Trouble Ticketing. Free.

6

Rather. mod_perl as a DSO is not going to be any fun. If you need mod_perl
as a DSO, you should probably use the fastcgi handler for the RT web interface.

Perfect. That has solved the problem! Thanks so much.

Jeremy.

Jesse

Nope. I’m using my own build of both. The most suspect thing in my setup
is that I’m building mod_perl as a DSO in order to preserve some other
complex apache mods.

Jeremy.

Are you guys using redhat’s builds of mod_perl and apache?

“Jeremy M. Smith” wrote:

Hello–
I’ve been happily using RT1 for almost three years. Today, I
decided to
make the jump to RT2. The installation went well, and everything from
my
RT1 instance imported perfectly. But there is some problem with any
RT2
form that uses the POST method instead of GET. With the main login
screen,
for instance, if I try to use the username and password form elements,
I
just get dumped back to the same screen. However, if I tag on
?user=xxxx&pass=yyyy to the base RT2 URL, I get in just fine. It seems
that most of the other forms use GET, so the system is functional, but
several of the administrative forms use POST.
I have checked the list archives. There have been a few threads
talking
about this problem, but no solutions aside from changing all of the
POSTs
in the code over to GETs, which seems clunky.
If I look in my webserver logs, it looks like the browser is
actually
sending two separate requests, almost as if the HTTP headers were
getting
messed up:
169.237.xxx.yyy - - [18/Oct/2001:13:39:01 -0700] “POST /rt2/
HTTP/1.1” 200
2071
169.237.xxx.yyy - - [18/Oct/2001:13:39:01 -0700]
“user=test&pass=testGET
/rt2/NoAuth/webrt.css HTTP/1.1” 200 1484
This problem is consistent across several different browsers,
including
Mozilla, IE 5.5, and Netscape 4.7.
I’m using RT 2.0.8, Apache 1.3.19, Perl 5.6.1, mod_perl 1.26 and
fresh
builds of all the required Perl modules (compiled from CPAN today) all
on
a RedHat 7.0 system.
Has anybody else seen this behaviour (and, hopefully, a solution)?
Thats two of us that have the same problem - but the only
difference is
that I am using Apache 1.3.20, and RedHat 7.1.
The only kludge I have is to change the posts to gets.
Any else any ideas ?
Greg

rt-users mailing list
rt-users@lists.fsck.com
http://lists.fsck.com/mailman/listinfo/rt-users

–
http://www.bestpractical.com/products/rt – Trouble Ticketing. Free

–

Jeremy M. Smith jeremy@ucdavis.edu
Technical Support Manager, Programmer
University of California, Davis
Languages and Literatures
Language Learning Center

rt-users mailing list
rt-users@lists.fsck.com
http://lists.fsck.com/mailman/listinfo/rt-users

Jeremy M. Smith jeremy@ucdavis.edu
Technical Support Manager, Programmer
University of California, Davis
Languages and Literatures
Language Learning Center

7

Jesse Vincent wrote:

Are you guys using redhat’s builds of mod_perl and apache?

All mine is custom complied.

Greg

8

Greg Cope wrote:

Jesse Vincent wrote:

Are you guys using redhat’s builds of mod_perl and apache?

All mine is custom complied.

oh, and statically as well

Greg