Re: non-blocking request to a url (via curl or file_get_contents or whatever)...
| From: | Jochem Maas | Date: | Sat, 20 Jan 2007 21:51:13 +0000 |
| Subject: | Re: non-blocking request to a url (via curl or file_get_contents or whatever)... | ||
| References: | 1 2 3 4 | Groups: | php.general |
| Request: | Send a blank email to php-general+get-247463@lists.php.net to get a copy of this message | ||
Roman Neuhauser wrote:
> # jochem@iamjochem.com / 2007-01-20 16:50:48 +0100:
>> Roman Neuhauser wrote:
>> 1. user stuffs things into shopping basket on [my] site (data stored in session)
>> 2. user goes to check out.
>> 3. user chooses online payment.
>> 4. user is redirected to online payment provider site
>> 5. user completes payment successfully
>> 6. online payment provider site contacts [my] site/server directly with transaction
>> status/details
>> 7. user is shown 'thank you' page on online payment provider site
>> 8. user is redirected back to [my] site and shown 'real' 'thank you'
>> page.
>
> That was perfect, thanks a lot!
>
>> normally in step 8 the webbug would be placed on the 'thank you' page,
>> but in this case the data needed to craft the webbug's url is no
>> longer available - the solution is to perform the request the the
>> webbug's url represents directly from my server during the
>> code that runs as a result of the request made by the online payment
>> provider in step 6.
>
> Is it important that the callback gets called synchronously?
dunno - definitely going to ask though!
in fact I feel stupid for not contemplating it myself,
I have a nasty suspicion that the script behind the url that the
tradedoubler webbug points to does stuff with the info the user's
browser would normally provide ...
in which case I would need to do something else - like generate the
webbug url at the point that I can (during order finalization) and store the
generated url in the relevant user's session and then use/place the webbug with
that url at the first opportunity to
> Is the
> order reconstructible from the callback url?
to some extent but I don't think that is relevant to my current little puzzle.
> If not I'd write
> a small script to fetch urls from a database table and feed them to
> wget or similar.
you've got me thinking about it from a totally different angle, and I've
got to understanding sockets/streams a little too! I now have enough ammo
to 'kill' the problem.
thank you very much for lending me your brain :-)
>