Hey Ryan,
How *exactly* did you locally do the build? If you had done the
install of your altered lift-paypal then you would certainly get a
compile error because the signature has changed. The new syntax should
be:
object MyIPN extends PaypalIPN {
def actions = {
case (Full(CompletedPayment), info, req) => // do something
}
}
The only exclusion would be if you had a implicit conversion to Box
PaypalTransactionStatus types that were unboxed.
Cheers, Tim
On Oct 9, 3:46 pm, Ryan Donahue <[email protected]> wrote:
> Tim,
>
> I locally changed the PaypalIPN.actions method return type to
> trait PaypalIPN {
> def actions: PartialFunction[(Box[PaypalTransactionStatus.Value],
> PayPalInfo, Req), Unit]
>
> }
>
> Apparently this does not cause any compilation errors for user
> implementing their own IPN handler as follows
> object MyIPN extends PaypalIPN {
> import PaypalTransactionStatus._
> def actions = {
> case (CompletedPayment, info, req) => // do something
> }
>
> }
>
> This is not good since I assume the result is that the case won't
> match anymore but we won't have a compilation error to tell us to
> change our code. Maybe I missed something, I am a scala newbie :)
>
> -Ryan
>
> On Oct 8, 3:57 pm, Timothy Perrett <[email protected]> wrote:
>
>
>
> > Ok cool, I'll take a look at this tomrrow all being well.
>
> > Thanks for the feedback
>
> > Cheers, Tim
>
> > Sent from my iPhone
>
> > On 8 Oct 2009, at 20:43, Ryan Donahue <[email protected]> wrote:
>
> > > I created the ticket:http://github.com/dpp/liftweb/issues/#issue/88
>
> > > I do receive the payment_status field for PDT. I bet in practice
> > > you will never receive a payment_status value other than Completed,
> > > because if the payment was not completed PayPal would not redirect
> > > the user's browser back to your PDT URL. However, I have not
> > > verified this and do check the payment status in my PDT code anyway
> > > (how could I verify that it will never happen?). So I would prefer
> > > that both were consistent, but just boxing the IPN payment status
> > > will be fine too :)
>
> > > On Thu, Oct 8, 2009 at 2:49 PM, Timothy Perrett <[email protected]
> > > > wrote:
> > > Im not married to the current API, so breaking changes are OK as
> > > there are only a handful of people using this code right now.
>
> > > To be honest, this whole situation just underlines the need for
> > > mocking in this module of lift... i've been meaning to do it since
> > > the beginning but just never got round to it and lack of general
> > > demand.
>
> > > Just about why they have a different signature... if memory serves
> > > that would be because PaypalTransactionStatus is not supplied for
> > > PDT. So whilst I see your point about them being consistent, im not
> > > sure there is value in having a Box parameter that would always be
> > > Empty? For IPN however, boxing sounds good to me.
>
> > > Should just be a case of:
>
> > > Box !! info.paymentStatus
>
> > > Would you be happy with that? If you could raise a ticket for this
> > > issue i'll patch it tomorrow and commit the code on a branch related
> > > to that ticket. One the code goes through review then it will make
> > > it into the master branch all being well.
>
> > > Cheers, Tim
>
> > > On 8 Oct 2009, at 18:52, Ryan Donahue wrote:
>
> > >> Yeah, I noticed my email got mangled.
>
> > >> It would make sense to me if PaypalIPN.actions and
> > >> PaypalPDT.pdtResponse were consistent.
>
> > >> trait PaypalPDT {
> > >> def pdtResponse: PartialFunction[(PayPalInfo, Req), LiftResponse]
> > >> }
>
> > >> trait PaypalIPN {
> > >> def actions: PartialFunction[(PayPalInfo, Req), Unit]
> > >> }
>
> > >> If matching the status is just too convenient to lose, then how
> > >> about this.
>
> > >> trait PaypalPDT {
> > >> def pdtResponse: PartialFunction[(Box
> > >> [PaypalTransactionStatus.Value], PayPalInfo, Req), LiftResponse]
> > >> }
>
> > >> trait PaypalIPN {
> > >> def actions: PartialFunction[(Box[PaypalTransactionStatus.Value],
> > >> PayPalInfo, Req), Unit]
> > >> }
>
> > >> Either would be an API breaking change for the paypal module, but
> > >> at least it'd be caught be the compiler. What do you think?
>
> > >> On Thu, Oct 8, 2009 at 1:33 PM, Timothy Perrett <[email protected]
> > >> > wrote:
>
> > >> The parameters were pretty mangled in your email... Which one would
> > >> you propose is more generic that the current status one we have?
> > >> Alternatively, it sounds to me like we might need to add some kind of
> > >> special case match statement.
>
> > >> Thoughts?
>
> > >> Cheers, tim
>
> > >> On 8 Oct 2009, at 13:14, Ryan Donahue wrote:
>
> > >> > Hi Tim,
>
> > >> > Looking at "Table 2. Summary of subscription variables" of the page
> > >> > you reference, payment_status is not included for messages with a
> > >> > txn_type of subscr_ signup, subscr_ cancel, subscr_ modify, or
> > >> > subscr_eot.
>
> > >> > This matches the results I see in testing with PayPal Sandbox.
> > >> > Neither subscr_signup or subscr_cancel include the payment_status.
> > >> > Below are the parameters from two messages I have received (they
> > >> > contain fake user info generated by PayPal Sandbox, so no security/
> > >> > privacy concerns).
>
> > >> > params from subscr_signup message:
> > >> > btn_id -> List(1070963), test_ipn -> List(1), charset -> List
> > >> > (windows-1252), address_country -> List(United States), item_name
> > >> ->
> > >> > List(Pro Account), recurring -> List(1), address_state -> List(CA),
> > >> > amount3 -> List(9.99), address_street -> List(1 Main St),
> > >> > verify_sign -
> > >> >> List(Ab4im9HUsk1pOL1dlUXJxEoipQcaAJQqV047JE2KENGFX4meCETo.KLt),
> > >> > address_status -> List(confirmed), period3 -> List(1 M),
> > >> subscr_date -
> > >> >> List(05:01:02 Oct 08, 2009 PDT), first_name -> List(Test),
> > >> > address_country_code -> List(US), address_city -> List(San Jose),
> > >> > mc_currency -> List(USD), mc_amount3 -> List(9.99),
> > >> > residence_country -
> > >> >> List(US), payer_email -> List([email protected]),
> > >> > address_name -> List(Test User), txn_type -> List(subscr_signup),
> > >> > payer_status -> List(verified), subscr_id -> List
> > >> > (S-8X87353621486401E), last_name -> List(User), payer_id -> List
> > >> > (JMM2RTLNU5RPJ), reattempt -> List(1), notify_version -> List(2.8),
> > >> > custom -> List(2), address_zip -> List(95131), receiver_email ->
> > >> List
> > >> > ([email protected]), business -> List
> > >> > ([email protected])
>
> > >> > params from subscr_cancel message:
> > >> > btn_id -> List(1070963), test_ipn -> List(1), charset -> List
> > >> > (windows-1252), address_country -> List(United States), item_name
> > >> ->
> > >> > List(Pro Account), recurring -> List(1), address_state -> List(CA),
> > >> > amount3 -> List(9.99), address_street -> List(1 Main St),
> > >> > verify_sign -
> > >> >> List(AvpXEetWAO5JsntihTfwE4HykmNlAsappICrYh6PaVgG3JJUYLtOnqVU),
> > >> > address_status -> List(confirmed), period3 -> List(1 M),
> > >> subscr_date -
> > >> >> List(05:06:19 Oct 08, 2009 PDT), first_name -> List(Test),
> > >> > address_country_code -> List(US), address_city -> List(San Jose),
> > >> > mc_currency -> List(USD), mc_amount3 -> List(9.99),
> > >> > residence_country -
> > >> >> List(US), payer_email -> List([email protected]),
> > >> > address_name -> List(Test User), txn_type -> List(subscr_cancel),
> > >> > payer_status -> List(verified), subscr_id -> List
> > >> > (S-8X87353621486401E), last_name -> List(User), payer_id -> List
> > >> > (JMM2RTLNU5RPJ), reattempt -> List(1), notify_version -> List(2.8),
> > >> > custom -> List(2), address_zip -> List(95131), receiver_email ->
> > >> List
> > >> > ([email protected]), business -> List
> > >> > ([email protected])
>
> > >> > Thanks,
>
> > >> > Ryan
>
> > >> > On Oct 7, 6:18 pm, Timothy Perrett <[email protected]> wrote:
> > >> >> Ryan,
>
> > >> >> The PayPal stuff is pretty much my baby - its awesome that your
> > >> using
> > >> >> it in production!
>
> > >> >> See what your saying - its been a while since I did anything with
> > >> >> PayPal; what fields *are* returned when the transaction is
> > >> canceled?
> > >> >> Or, specifically, the question is what behaviour do you need?
>
> > >> >> Lets start there and then I can work on a solution - it was my
> > >> >> understanding that all responses from paypal had the
> > >> payment_status
> > >> >> field; im using the reference here:http://is.gd/43po4canyou
> > >> detail
> > >> >> the type of transaction you've been seeing this behaviour on?
>
> > >> >> Cheers, Tim
>
> > >> >> On 7 Oct 2009, at 20:56, Ryan Donahue wrote:
>
> > >> >>> I'm receiving IPN updates from PayPal when a subcription is
> > >> created
> > >> >>> and when it is canceled. However, the cancel message never make
> > >> >>> it to
> > >> >>> my actions method. Apparently, the cancel message does not
> > >> include
> > >> >>> the payment_status. I think this results in the message being
> > >> >>> dropped
> > >> >>> on the floor because of the for-comprehension on line 458 in
> > >> >>> Paypal.scala. Would appreciate any help. By the way, we've been
> > >> >>> using the paypal module in production for a couple months now (a
> > >> >>> very
> > >> >>> small app for selling a partner-branded version of our screen
> > >> >>> recorder).
>
> > >> >>> Thanks,
>
> > >> >>> Ryan
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Lift" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/liftweb?hl=en
-~----------~----~----~----~------~----~------~--~---
