Concerning 414 Request-URI Too Long

["Mark R. Diggory" <md...@latte.harvard.edu>]

Is there any way to increase the Request-URI buffer so it is larger? We
need to be able to handle larger GET requests...

-Mark Diggory

Re: Concerning 414 Request-URI Too Long

["Craig R. McClanahan" <Cr...@eng.sun.com>]

"Christopher K. St. John" wrote:

>
>  Putting my money where my mouth is: Would you
> accept a change that made the exact limit a config
> parameter? The current limit isn't just obscure
> and hardcoded, it's different than Apache...
>

Sure ... I have no problem with that if we can make sure performance for *other*
requests is not adversely affected (given sufficient memory in your server, of
course).

>
> -cks
>

Craig

Re: Concerning 414 Request-URI Too Long

["Mark R. Diggory" <md...@latte.harvard.edu>]

We're dealing with statistical proceedures on large datasets and trying to capture the
state of the current statistical "proceedure" and "data variables used" in a bookmark.
The goal is not to have to store this "citation" info on the server, just the data and
the statistical engine to generate the results. The bookmark tells the engine how to
recreate the desired statistics/visual output. Given that we need to tell the engine
which variables to use (usually >100) we tend to create large url's the test the
limits of browsers and servers. I don't think its poor design. We are trying to create
url's/bookmark's that are persistent and may be used for citations in electronic
journals. There's alot of information we're trying to capture the state of and store
in the bookmark. otherwise we're stuck providing storage and housecleaning on these
"citations" on the server side.

Apache and Tomcat are excellent products, I would be here asking these questions if I
thought otherwise. I wasn't looking to create arguments concerning compliancy.  After
searching the documentation and not finding anything on uri limits, the next logical
place to ask seemed to be here. Obviously arbitrarily long url's could be dangerous
for security and performance. I was just looking for for the possibility of adjustable
limits in the currently released product. If there are none, as it would seem, then I
thank you for your time and I continue on my journey.

-Mark Diggory


"Christopher K. St. John" wrote:

> "Craig R. McClanahan" wrote:
> >
> > "Christopher K. St. John" wrote:
> >
> > >  Doesn't  that mean Tomcat is merely "conditionally
> > > complaint" because it doesn't satisfy all the SHOULDs?
> > >
> >
> > According to the HTTP/1.1 definition of compliance, you can make that case if
> > you want.  IMHO anyone who needs URIs that long has a bad application design and
> > should fix it anyway.
> >
>
>  I suppose it's too late in the history of HTTP
> to make any sort of productive stand on the issue,
> but part of the point of implementing the whole
> protocol fully is that someday, new applications
> may want to take advantage of a feature and not
> be able to, since that bit of the spec wasn't
> widely implemented.
>
>  It's a "libraries are judged not by how well
> they do what they were designed to do, but how
> well they adapt to doing things that the initial
> implementers never thought of" sort of thing.
>
>  Putting my money where my mouth is: Would you
> accept a change that made the exact limit a config
> parameter? The current limit isn't just obscure
> and hardcoded, it's different than Apache...
>
> -cks
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: general-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: general-help@jakarta.apache.org

Re: Concerning 414 Request-URI Too Long

["Christopher K. St. John" <cs...@quik.com>]

"Craig R. McClanahan" wrote:
> 
> "Christopher K. St. John" wrote:
> 
> >  Doesn't  that mean Tomcat is merely "conditionally
> > complaint" because it doesn't satisfy all the SHOULDs?
> >
> 
> According to the HTTP/1.1 definition of compliance, you can make that case if
> you want.  IMHO anyone who needs URIs that long has a bad application design and
> should fix it anyway.
> 

 I suppose it's too late in the history of HTTP
to make any sort of productive stand on the issue,
but part of the point of implementing the whole
protocol fully is that someday, new applications
may want to take advantage of a feature and not
be able to, since that bit of the spec wasn't
widely implemented.

 It's a "libraries are judged not by how well
they do what they were designed to do, but how
well they adapt to doing things that the initial
implementers never thought of" sort of thing.

 Putting my money where my mouth is: Would you
accept a change that made the exact limit a config
parameter? The current limit isn't just obscure
and hardcoded, it's different than Apache...


-cks

Re: Concerning 414 Request-URI Too Long

["Craig R. McClanahan" <Cr...@eng.sun.com>]

"Christopher K. St. John" wrote:

> "Mark R. Diggory" wrote:
> >
> > Christopher,
> >
> > I have 3.2, we are on a development crunch and I don't expect to begin
> > testing 4.0 until I've met my deadline.
> > ...
> > Currently, I am hoping for a configuration solution instead of a coding
> > solution.
> >
>
>  I was wrong. 4.0 does this thing where it starts at
> 64 bytes and escalates up to 4096, but it doesn't
> handle arbitrary lengths (which would be problematic
> for obvious reasons), and the 4096 is hardcoded
> as a public final static, so you can't change it
> without recompiling.
>
>  Doesn't  that mean Tomcat is merely "conditionally
> complaint" because it doesn't satisfy all the SHOULDs?
>

According to the HTTP/1.1 definition of compliance, you can make that case if
you want.  IMHO anyone who needs URIs that long has a bad application design and
should fix it anyway.

NOTE:  It matters not at all how long a URI can be for Tomcat, if you've got a
proxy in between that has a lower limit ... and this is a *very* common scenario
in corporate intranets behind a firewall.

Or course, you can always fix and recompile if you disagree ... open source is
nice that way.

>
> -cks
>

Craig McClanahan

Re: Concerning 414 Request-URI Too Long

["Christopher K. St. John" <cs...@quik.com>]

"Mark R. Diggory" wrote:
> 
> Christopher,
> 
> I have 3.2, we are on a development crunch and I don't expect to begin
> testing 4.0 until I've met my deadline.
> ...
> Currently, I am hoping for a configuration solution instead of a coding
> solution.
> 

 I was wrong. 4.0 does this thing where it starts at
64 bytes and escalates up to 4096, but it doesn't
handle arbitrary lengths (which would be problematic
for obvious reasons), and the 4096 is hardcoded
as a public final static, so you can't change it
without recompiling.

 Doesn't  that mean Tomcat is merely "conditionally
complaint" because it doesn't satisfy all the SHOULDs? 


-cks

Re: Concerning 414 Request-URI Too Long

[Jon Stevens <jo...@latchkey.com>]

on 12/11/2000 2:25 PM, "Mark R. Diggory" <md...@latte.harvard.edu> wrote:

> Currently, I am hoping for a configuration solution instead of a coding
> solution.

Great! Send a patch. :-)

-jon

Re: Concerning 414 Request-URI Too Long

["Mark R. Diggory" <md...@latte.harvard.edu>]

Christopher,

I have 3.2, we are on a development crunch and I don't expect to begin
testing 4.0 until I've met my deadline. 

Jon, pardon me, I didn't mean to sound like I was flaming. I was just
getting responses that were not answering the question I asked, so I
made it stand out more. 

Currently, I am hoping for a configuration solution instead of a coding
solution.

Honk, honk

-Mark Diggory


"Christopher K. St. John" wrote:
> 
> Jon Stevens wrote:
> >
> > on 12/11/2000 1:22 PM, "Mark R. Diggory" <md...@latte.harvard.edu> wrote:
> >
> > > Again, "IS THERE ANYWAY" to allow for a larger GET URI in Tomcat?
> >
> > YOU HAVE THE SOURCE CODE!
> >
> 
>  OTOH, requring end-users to also be coders is probably
> a bit unrealistic. A more helpful answer might be along
> the lines of "4.0 doesn't appear to have the limit anymore,
> what version are you using?"
> 
> -cks
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: general-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: general-help@jakarta.apache.org

Re: Concerning 414 Request-URI Too Long

["Christopher K. St. John" <cs...@quik.com>]

Jon Stevens wrote:
> 
> on 12/11/2000 1:22 PM, "Mark R. Diggory" <md...@latte.harvard.edu> wrote:
> 
> > Again, "IS THERE ANYWAY" to allow for a larger GET URI in Tomcat?
> 
> YOU HAVE THE SOURCE CODE!
> 

 OTOH, requring end-users to also be coders is probably
a bit unrealistic. A more helpful answer might be along
the lines of "4.0 doesn't appear to have the limit anymore,
what version are you using?"


-cks

Re: Concerning 414 Request-URI Too Long

[Jon Stevens <jo...@latchkey.com>]

on 12/11/2000 1:22 PM, "Mark R. Diggory" <md...@latte.harvard.edu> wrote:

> Again, "IS THERE ANYWAY" to allow for a larger GET URI in Tomcat?

YOU HAVE THE SOURCE CODE!

-jon

-- 
Honk if you love peace and quiet.

Re: Concerning 414 Request-URI Too Long

["Mark R. Diggory" <md...@latte.harvard.edu>]

It appears to be a point of conflict. I've come across something similar
in the archive with no resolution. Someone argued that Apache and
Netscape Netserver allow for differing length URi's and both are larger
than Tomcat's 1024 byte buffer. 

Again, "IS THERE ANYWAY" to allow for a larger GET URI in Tomcat? A
setting in the Server.xml or something in some properties file
somewhere? If not, this would be beneficial in future versions.

-Mark Diggory

"Christopher K. St. John" wrote:
> 
> Adam Goldband wrote:
> >
> > The limit on the GET request are imposed as part of the RFC on HTTP
> > requests.
> > See RFC 1945:
> >
> > http://www.cis.ohio-state.edu/htbin/rfc/rfc1945.html
> >
> 
>  HTTP1.1 (rfc2616) sez:
> 
>     The HTTP protocol does not place any a priori limit on the length of
>    a URI. Servers MUST be able to handle the URI of any resource they
>    serve, and SHOULD be able to handle URIs of unbounded length if they
>    provide GET-based forms that could generate such URIs.
> 
>  It goes on to note that even if servers can handle ultra
> long URI's, some clients might not...
> 
> -cks
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: general-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: general-help@jakarta.apache.org

Re: Concerning 414 Request-URI Too Long

["Christopher K. St. John" <cs...@quik.com>]

Adam Goldband wrote:
> 
> The limit on the GET request are imposed as part of the RFC on HTTP
> requests.
> See RFC 1945:
> 
> http://www.cis.ohio-state.edu/htbin/rfc/rfc1945.html
> 

 HTTP1.1 (rfc2616) sez:

    The HTTP protocol does not place any a priori limit on the length of
   a URI. Servers MUST be able to handle the URI of any resource they
   serve, and SHOULD be able to handle URIs of unbounded length if they
   provide GET-based forms that could generate such URIs.

 It goes on to note that even if servers can handle ultra
long URI's, some clients might not...


-cks

Re: Concerning 414 Request-URI Too Long

[Adam Goldband <ad...@cnet.com>]

The limit on the GET request are imposed as part of the RFC on HTTP
requests.
See RFC 1945:

http://www.cis.ohio-state.edu/htbin/rfc/rfc1945.html


If you change your GET's to POST's, though, you will find that you have much
more size to your buffer . . . unless you need to be able to bookmark these
requests . . .

adam

----- Original Message -----
From: "Mark R. Diggory" <md...@latte.harvard.edu>
To: <ge...@jakarta.apache.org>
Sent: Monday, December 11, 2000 12:07 PM
Subject: Concerning 414 Request-URI Too Long


> Is there any way to increase the Request-URI buffer so it is larger? We
> need to be able to handle larger GET requests...
>
> -Mark Diggory
>


----------------------------------------------------------------------------
----


> ---------------------------------------------------------------------
> To unsubscribe, e-mail: general-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: general-help@jakarta.apache.org