OK, I think I have stumbled on a solution. Which I thought would be useful
to share here.
Playing with request.env, I noted that this too was Storage.
Hence: request.env.query_string appears to give me exactly what I need.
i.e. it is essentially the "raw" string of vars. Unaltered and in the
original order, just as PayPal wants to see.
So, anyone planning to deploy PayPal's IPN, might like to make a note.
This doesn't appear to be detailed anywhere ?? If it is, then I have
overlooked a useful source of API information (that might have led me this
answer a lot earlier and not wasted forum space). If anyone knows where I
would have found this detailed, I would really appreciate the URLs
please......
Thanks
Simon
On Thursday, July 18, 2013 5:46:23 PM UTC+1, SimonD wrote:
>
> Hello, I would be sure that someone has come across this already. I hope
> so.
> I am writing a handler for PayPal IPN. Actually, a little easier that I
> thought it would be except for one issue: -
>
> request.vars creates a Storage object, but the key:value pairs are not in
> the same sequence as they are in the URL. It almost seems to be a
> randomised order.
> For 99.9% of occasions this is OK. However, as part of the IPN handler
> authentication, PayPal expects to receive the URL (var) string back exactly
> as they posted it to my application, specifically with the vars in the same
> order. Their documentation re-states this requirement in multiple places !.
> I suspect (although I don't know), that PayPal is simply matching the
> return var string to the one it posts. i.e. not actually wasting server
> cycles with intelligent var matching :-(
> I am also not on a LIVE server so, although PayPal have a test sandbox, it
> is not a lot of use at this point.
>
> I cannot identify any other function in web2py to pull the raw URL string.
> I think the only function is request.vars.
> I note that request.args has an index method (i.e. request.args(i) ) so
> that the args can be pulled in sequence. But request.vars does not have
> this method.
>
> To test my random order theory, here is a test function()
>
> import urllib
> def sample():
> #sample of ipn_handler code
> getvars=request.vars
> test1=urllib.urlencode(dict(getvars))
>
> getvars=request.vars
> test2=''
> for item in getvars:
> test2=CAT(test2,'&',item,'=',getvars[item])
>
> test3=request.vars
>
> return dict(test1=test1,test2=test2,test3=test3)
>
> and the view (extract)....
> test1: {{=test1}}<br />
> test2: {{=test2}}<br />
> test3: {{=test3}}<br />
>
> So, if my url vars are as follows:-
> ?a=11&b=22&c=33&d=44&e=55&f=66&g=77&h=88
>
> ...this is what the view shows.
> test1: a=11&c=33&b=22&e=55&d=44&g=77&f=66&h=88
> test2: &a=11&c=33&b=22&e=55&d=44&g=77&f=66&h=88
> test3: <Storage {'a': '11', 'c': '33', 'b': '22', 'e': '55', 'd': '44',
> 'g': '77', 'f': '66', 'h': '88'}>
>
> You can see that I have testing two ways to create a return var string
> (test1 and test2), and also shown the Storage object (test3)
> test1, test2 and test3 show an identical sequence. But it is different to
> the actual posted URL var sequence.
> BTW - the actual IPN string is not determiistic, so it is not like I can
> manually build a return string.
>
> Thanks!
> Simon
>
>
>
> I note that on web2pyslices, there are (I see) two slices (for PayPal IPN)
> that also appear to generate a randomised return string. According to the
> PayPal docs, that string will be rejected.
>
> So, what on earth am I doing wrong? Or perhaps there is another way to
> pull the vars from the URL? (request.url doesn't pass the vars)
> Or has someone a different way to recreate the URL for PayPal IPN?
>
> Thanks
> Simon
>
>
--
---
You received this message because you are subscribed to the Google Groups
"web2py-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to web2py+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
