[digikam] [Bug 374225] Add possibility to remove face identities by removing tag and remove person tags but preserving the tags themselves [patch]

Simon bugzilla_noreply at kde.org
Wed Jan 11 21:20:21 GMT 2017


--- Comment #11 from Simon <freisim93 at gmail.com> ---
Created attachment 103361
  --> https://bugs.kde.org/attachment.cgi?id=103361&action=edit
Mario's patch with changes and minus my patch

I should have written that I don't consider my patch suitable anymore, for the
reasons you stated, because it prompts the user for confirmation for every
deleted tag when multiple face tags are deleted and it doesn't remove the
actual tag regions from faces.

The attached patch removes my part, as Mario's part is self contained and mine
not ready to be pushed. I also introduced the following changes:
The slotMultipleFaceTagDel and slotFaceTagDelete methods are mostly redundant,
so I replaced the latter with a call to the former. This introduces a tiny
overhead, which is IMO worth it for less duplicity and thus easier maintenance.
I also did some purely cosmetic changes (mostly line breaks). If you actually
prefer long lines, please tell me and I will stop doing this in the future.

Some conceptual questions:
The naming "Delete person tag" could be misunderstood as actually deleting the
tag, not just its "affiliation" to a face. Something like "Delete face
(identity)"/"Remove person from tag" would IMO be clearer, but it doesn't
completely convince me either. I think it is necessary to clarify this, any
To implement the part I already implemented (incompletely): My preferred option
would be to extend the existing function that removes a tag to also remove the
face identity and regions when the deleted tag is a person tag, also if the
people sidebar is not active. This seems the cleanest way, as otherwise we have
information "hanging around" in the db and metadata, that is not accessible
from the UI (the user can still keep this stuff). The downside: It adds even
more potential popup questions the user has to answer. Opinions?

You are receiving this mail because:
You are the assignee for the bug.

More information about the Digikam-devel mailing list