On Thu, Dec 12, 2019 at 07:38:37PM +0100, Stefan Sperling wrote:
> On Thu, Dec 12, 2019 at 07:02:08PM +0100, Stefan Sperling wrote:
> 
> A similar situation exists in your add -R patch and I failed  to spot it.
> The addition code is also calling got_fileindex_entry_get() with one path
> and get_file_status() with the obtained file index entry but with an
> unrelated path.
> 
> Here is how I would solve that case. The status walk already takes care of
> figuring out what is a file and what is a directory. Which means that
> schedule_addition() will only be called on files in the first place.
> So the 'ondisk_path' in struct schedule_addition_args is simply not needed :)
> Perhaps this helps with figuring out how to improve your deletion patch?

Ok. That makes sense. I'll look at the remove patch some more. Thanks.

-- 

Tracey Emery