URL lengths

Paul (Paul.Wain@brunel.ac.uk)
Thu, 17 Mar 1994 16:57:05 --100


Hi,

I probably shouldnt post this one here but until I remember Lou's
address (oops!) I will since it seems to apply to more than one package
anyway :)

Are there any (known) limits on the lenghts of URLs? Im working on
something right now that needs to encrypt a LOT of information into a
302 redirect and it appears to cause problems if the line gets too
long.

The redirect is somewhat nasty since it needs to contain a LOT of path
information. To see what I mean I guess the best thing to do is look at:

http://http2.brunel.ac.uk:8080/paul/X.500/

Submit this form without entering a search pattern (a valid request in
the case of the way that the program was designed) and you submit some
data to be resolved into a redirect to:

http://echo.brunel.ac.uk:4040/

This then returns another form (eventually the 1st form wont be there
but for now this is the way it works) configured from the 1st one.

However:

a) Mosaic wont redirect, and returns the page that is "Forward" in its
history, (only if you have a forward history mind otherwise it does
nothing) and

b) Lynx tries to render the page that has the redirect in it.

c) PC Mosaic barfs at the redirect despite claiming to be HTTP/1.0
compliant :)

I think then that I am running out of space in a buffer somewhere. The
URL is about 270 chars long at this point so Im thinking that somewhere
down the line the clients have a char buffer[256] or something :)

Anyway, ideas would be appreciated. Reasonably sneaky programming has
enabled a work around locally but I dont think it would work if the work
I am doing was done at another site.

For a look at what it is I am doing, (cos I am sure you are all curious
now that I mention it) an alpha1 is located on:

http://echo.brunel.ac.uk:4040/

and the alpha2 (which is the one causing problems) is through:

http://http2.brunel.ac.uk:8080/paul/X.500

The 2nd one may not always be avaliable though since it is running under
dbx and has at least one memory leak that is causing problems.

Anyway clues as to if I am breaking a wwwlib or something that the
clients are doing internally for redirects would be nice.

(BTW I did scan the HTTP/1.0 and HTML/1.0 specs and couldnt find any
limits on URL lengths... is that right?)

Many thanks,

Paul
.-------------------------------------------------------------------------.
| Paul S. Wain, (X.500 Project Engineer and WWW/HTTP chappie), |
|-------------------------------------------------------------------------+
| Computer Centre, Brunel University, Uxbridge, Middx., UB8 3PH, ENGLAND. |
| VOICE: +44 895 274000 extn 2391 EMAIL: Paul.Wain@brunel.ac.uk |
| http://http1.brunel.ac.uk:8080/~ccsrpsw/ |
`-------------------------------------------------------------------------'