dont cache if model::$isCachable == false

[hxnk]

When calling model::all(), the $isCachable property on the model wasn't checked, resulting in model:all() results to be cached, even when $isCachable was set to false.

This PR attempts to fix that :)

[coveralls]

Coverage decreased (-0.3%) to 99.392% when pulling 2e8b2d0 on hxnk:respect-isCachable into 7a7bed0 on GeneaLabs:master.

[mikebronner]

The problem is that isCachable can never be set for the all() method, since setting isCachable is set through a query scope, which returns a Builder object, which doesn't have an all method. Consider that the following code throws an exception:

$model->disableCache()->all();

Unfortunately the only way to disable caching on all() queries is to turn it off in config at this point, or change to get(). Perhaps you have some thoughts on how we can disable the cache on the model? In the past, I've had a method on the model, instead of a scope, and that had more problems than it solved, so I moved to scope.

[mikebronner]

Hi @hxnk, thank you for submitting this PR! :) I left a comment explaining why this code won't work at this point. Please let me know if you have additional ideas. :)

[hxnk]

Hi @mikebronner, what do you think of this solution? It's a bit hacky, but it works :)

• disableCache() temporarily disables cache in the config and returns a QueryOrModelCaller object, that is able to direct chained calls to the right objects (all() to Builder, rest to Model). This allows using a query scope and still chain the all() call, which would normally not be possible since a Builder object would be returned
• in the QueryOrModelCaller object cache setting in the config is set back to its original state, so that subsequent queries aren't affected

[mikebronner]

Hi @hxnk, I initially thought that might work as well, but then realized that any global method of disabling cache would cause unexpected behavior for other possible simultaneous request. To avoid uninterrupted side-effects, we will have to find a solution that works on the model instance.

[mikebronner]

Hi @hxnk -- I was able to come up with a solution! :) Thanks for keeping the discussion going around this challenge and prompting me to think about different ways of achieving this!