(This question was previously raised on IRC with no response.) Issue #11421 <https://redmine.darktable.org/issues/11421> describes a situation where image and XMP files are shared between clients and a reliance on the write_timestamp noted in the local database and XMP file to determine XMP file freshness can result in unnecessary XMP writes and later reconciliation.
Would it make sense to address this issue using a combination of write_timestamp and XMP file hash? The idea would to update the timestamp in the local database to a newer one if the file hash were unchanged and not treat the files as having been updated. If so: 1. Is there a preference for a particular hash function? Some hash functions yield shorter output with less information, which may be helpful for database size. 2. Is there a preference regarding computation (e.g., read back from disk after write or write to a buffer and keep the hash)? 3. Are there any preferences with regard to third-party libraries (e.g., OpenSSL or similar)? Thank you in advance. ___________________________________________________________________________ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
