Full index support for metadata validation error
Currently, metadata validation error messages are only indexed if the field does not exceed a certain max length. If the file contains a longer error message, the IE fails indexing and lands in the exception queue, leading to no meaningful way to find them other than by their SIP / IE ID.
This becomes problematic, when validation error messages are exceptionally long - either per design or per error in embedded third party tools. A current example for this exists through a jhove misbehavior, which is documented as an issue here: https://github.com/openpreserve/jhove/issues/165
Removing the validation error message from the Index Queue alltogether is not an option, as this would prohibit targeted set building / preservation action. Truncating them would be an option.
It should be ensured that objects do not fail indexing due to long validation error messages.
The indexing limitation was removed and will be available as part of Rosetta’s next version (5.2).
The indexing limitation was removed and will be available as part of Rosetta's next version (5.2).
Jan Hutar commented
agreed, also we might want to look at this issue from broader perspective of other indexing issues with too long "fields" and try to find kind of general solution