World: r3wp

I've updated my cheyenne.exe binary to use my prot-http.r so that 
you can do PUT, DELETE, GET with read/custom etc.

Graham, not only you, but that's a good thing. :-) Cheyenne would 
have never become as solid as it is today without your and others 
help.

Henrik: on Windows, you have a help message when the mouse pointer 
is over Cheyenne's tray icon. From RSP script, just use :

>> p: open http://localhost
>> p/locals/headers/server
== "Cheyenne/0.9.20"

ok, thanks

Will, are you familiar with how to use Cheyenne's logging functions 
from within a rsp script?

not sure what you mean, rsp-log which is deprecated in latest version?

found it .. can use ?? to log to the trace.log

This is really really strange.

I have a rsp page called creategoogledoc.rsp ... I click on it once, 
and the I get my document created.

Click on it again and the rsp script is not executed ( looking from 
the trace ) but I still get taken to the googledoc page where I previousl 
y created the document.

It's as though the browser is caching my request and executing it 
locally without hitting the server

Arrrggg.... it's all caused by Chrome.  Switching to FF and the problem 
goes away!

And here I was looking for  some type of caching occuring in Cheyenne 
or my RSP pages

I wonder if there is anything on the cheyenne side that can stop 
this behaviour?

rsp pages should not be cached by default by any browser. Do you 
have firebug in FF can you look at the header fo the response an 
see if there is any difference between the two requests? wireshark 
or tcpflow can be used to look at headers as well

I was having similar cashing problems with Chrome...I could not getr 
it to display a changed/updated rsp page. I had to close out.

interesting

If this only occurs with Chrome and RSP ... maybe there is something 
that can be done.

Google Chrome has some badly broken logic when it comes to obeying 
cache-control headers.

Got a link to this ?

Here's someone else banging their head against chrome's caching logic:

http://stackoverflow.com/questions/559984/getting-confused-about-expires-headers-when-testing-in-chrome

That appears to be saying that Chrome is not caching ....

here the problem appears to be that Chrome is heavily caching.

Lots of other people seeing overheavy  Chrome caching, eg:

http://www.google.com/support/forum/p/Chrome/thread?tid=77a9194da602bc00&hl=en

I've created a youtube video showing what is happening http://www.youtube.com/watch?v=ANgSrOuDncE

I'm using BB FlashBack Express 2 because it can upload directly to 
youtube ..and it seems to be free :)

Good idea there .. I'll try an incognito window to see if that helps.

Sadly no .. it does actually create the document the first time, 
but then brings up the same document again the second time.   So, 
basically the same behavior as cognito

I think I may be able to get around this issue by tagging the request 
with a dummy time parameter.

That way chrome will not think it's the same idempotent request

My server traces show that Chrome is not actually visiting the pages 
...

Try by setting a different ETag header for each document: response/set-header 
'Etag checksum document

You can use a timestamp or doc ID for the ETag value, it just has 
to be unique for each resource.

in the url??

Nope, in the RSP script.

http://en.wikipedia.org/wiki/HTTP_ETag

Graham, can you paste here the response headers of the first request 
please

well changing the url by putting a timestamp fixes the problem partially.

HTTP/1.1 200 OK
Content-Type: text/html; charset=UTF-8
Date: Wed, 19 Aug 2009 07:41:04 GMT
Expires: Wed, 19 Aug 2009 07:41:04 GMT
Cache-Control: private, max-age=0
X-Content-Type-Options: nosniff
Server: GFE/2.0
Via: 1.1 bc3
Content-Length: 0
Connection: Keep-Alive
Set-Cookie: cookies here ... 

Set-Cookie: user=; Expires=Tue, 18-Aug-2009 07:41:04 GMT; Path=/; 
HttpOnly

Set-Cookie: login=; Expires=Tue, 18-Aug-2009 07:41:04 GMT; Path=/; 
HttpOnly

Reading about etags now ...

It's not the document that needs to send back the etag as it never 
even gets that far.

So, perhaps I need to set the etag in the page that lists links?

why do u have Content-Length: 0 ?
this is not valid Set-Cookie: cookies here ...

I scrubbed the cookies from the response

this is returned by googledocs

are u reverse proxying ?

no proxies

made the etag settings ... not helping so going to do a wireshark 
trace to see if I am setting it

what if you add
response/set-header 'Expires "-1"

it's caching the whole page ... so it won't see the etags

headers you pasted are not coming from cheyenne but from google server, 
you can't control them

just to understand, you generate a file and u upload to googledocs, 
then requests it from google server?