function quickfile_link() generates an improper download URL

I found an example where function quickfile_paypal_ipn() never did insert a record for one of my test transactions, so the cause may be something other than a timing issue.

Sorry (red faced)

JM

(still red faced)

I just learned that changing the title on an issue reply changes it for the whole issue.

JM

Last post for me tonight. (I'm clearly too tired to think as carefully as I need to.)

It turns out the reason no transaction record was posted in my #2 reply is because I forgot to put my test site back online before running another test case.

Results from my earlier tests all point to a timing issue. I am still convinced this is a timing issue.

JM

... now better rested and fully alert :)

Thanks alexis. I am happy to help with testing, and I can also provide you with access to my test environment.

Quickfile is a *much* better solution than what I'm currently using , so I'm very eager to assist.

Here are some more details that may be helpful:

I have a test site setup in a shared hosting envornment. The web server and database server are on separate machines. The test environment is connected to the Paypal sandbox and running with Quickfile's Paypal sandbox mode enabled.

The bad URL results seem to happen sporadically, but it's usually pretty easy to reproduce the problem by running the same scenario a few times. When the problem appears, I am always able to get correct results by telling my browser to refresh a few seconds later. From what I've seen in the code, this "refresh" observation is key to my conclusion that the problem is caused by a timing issue.

** If you're interested, I'd be happy to write some code that uses the payment processor's transaction ID in the download link (and could be used as part of the fix).

OFF TOPIC -- I plan to modify the standard release to allow users 3 downloads of the same file within 30 days. This is why I'm kind of pushing to use the payment processor's transaction ID to make the download link a little more secure. :)

Since I'm not yet using this is in production, I'm downgrading the priority to normal.

Thanks again,

JM

Thanks alexis

It's been pretty easy to reproduce in my environment this week. I think load and configuration of servers has a lot to do with it, as these factors affect execution times. My test setup is in a cheap shared host environment, so that probably helps with reproducing the problem :)

I'm available 12/20 & 12/21 to assist with testing. Sending an email via the contact form in my profile is the quickest way to get my attention.

JM