Questions about environment dependencies

[goldenapples]

Hi, I'm just getting started playing with this library. One thing I've noticed is that in my local development environment (a more or less stock VIP Quickstart Vagrant box on a Macbook Pro), the post to admin-post.php always fails with the 0.01s timeout in the plugin.

Is there a reason why this needs to be set so low? The request is non-blocking anyways, Is there any reason a default timeout can't work?

In addition, the only way I could make the admin-post callback complete was by adding @ignore_user_abort( true ); to my local admin-post.php (or the equivalent value in php.ini, though that's even sketchier...). Is this a requirement for the plugin to work, and if so, should it be documented somewhere?

[rayrutjes]

Currently also having issues debuging the remote calls. Locally I have every request failing but the actual tasks are correctly run. That said, on some other environments, the task will not run, but the errors remain the same.

I agree with @goldenapples , we should try to document the requirements for the "async" trick to work. Currently I do not feel it is very reliable.

Some points that needs to be addressed:

• The host must be reachable from inside the machine
• Explain the timeout, and if it is ok always failing
• Any SSL related requirements?
• Explain differences between auth. users and not, and also give some examples
• Other?

Great work though! Keep it up!

[r-a-y]

the post to admin-post.php always fails with the 0.01s timeout in the plugin.

0.01 is the same timeout as used when WP triggers cron in spawn_cron(). This timeout value has been talked about on WP Trac a few times, most notably:
https://core.trac.wordpress.org/ticket/18738

There probably should be a filter for the arguments in wp_remote_post() similar to WP Cron. See #12.