Introduce the general idea of overwrite NA/TRUE/FALSE #266
Conversation
In previous refactoring commit, I stopped checking whether both `path` and `parents` were specified in `drive_upload()`. I claim this is incredibly rare and not worth the code and headspace to maintain. Let the behaviour be undefined / undescribed and, if it is wrong or surprising and actually matters, someone will speak up.
Tests added in the next commit revealed that, for filepath input to `drive_rm()`, we tried to delete each file n times, where n = the number of filepaths for that file. n can be greater than 1 if a file has multiple parents.
I am removing checks and tests around unusual parent situations elsewhere. I intend to send people to drive_update() for all weird parent folder situations. So I want an explicit test that you can add a parent folder this way.
Revisit this later with lifecycle
Closes #230
jennybc
changed the title
|
@LucyMcGowan This is the biggest piece of work in googledrive since its release (other than the adoption of gargle, which didn't really add or change a Drive feature per se). I still need to add tests for this and make sure And I can fly solo here. But if you are in a position to look this over, it is always great to have another set of eyeballs. Even if you just read and comment, that's welcome. |
Refine tests
But I have seen that the presence of files on Drive, in the trash, can affect tests. Tests should technically leave the trash the way the found it and be "trash-aware". But sometimes it's easiest to just clean it out.
Truly closes #230
|
I took a quick look, and this looks awesome! The only thing I noticed is we already had an overwrite parameter for |
|
Thanks @LucyMcGowan, that's a good point. I added a "local" qualifier there, to serve as a reminder of where the overwrite could occur. |
Closes #230
I decided that, if
overwritehappens anywhere (such as indrive_upload(), it should happen for the entire set of functions that create new files.drive_put()is also new.