Changed ActiveFedora::File#persisted_size to simply return 0 if we're lo...

[afred]

...oking at a new record, rather than attempt a head request on the Ldp::Resource::BinarySource object.

[coveralls]

Coverage increased (+0.0%) when pulling 5c43458 on afred:change-af-file-persisted-size into 2db7e03 on projecthydra:fedora-4.

[afred]

@awead @jcoyne This should avoid an Ldp::NotFound error when trying to do a head request on the Ldp::Resource::BinarySource object in ActiveFedora::File#content when the ActiveFedora::File object is new.

There is still a case where #has_content? will be false for new ActiveFedora::File objects that have something in #content that does not respond to a :size method.

One solution, would be to have #dirty_size do more thorough duck typing on #content looking for more than just :size but any other method that could provide an accurate value. I don't know what those methods would be though. Do we have a list of all the types of object that we can expect ActiveFedora::File#content to be?

[afred]

👍 for #persisted_size to return nil if the record is not persisted. I'll push the change here momentarily.

Do you have any comment on the last bit of my comment above?

[awead]

@afred content should always respond to .size. If it doesn't, it should raise an error. Are you seeing a case where is doesn't?

[coveralls]

Coverage remained the same when pulling 2751b25 on afred:change-af-file-persisted-size into 2db7e03 on projecthydra:fedora-4.

[afred]

@awead no case that i can think of. Just wanted to be sure.

[afred]

oh damn... i forgot to change the test expectation methinks. Stand by...

[coveralls]

Coverage remained the same when pulling e468379 on afred:change-af-file-persisted-size into 2db7e03 on projecthydra:fedora-4.