While removing mptt dependency papermerge.search.tasks.update_index was changed: instead of updating index for each individual instance - the whole index is updated (similarly to invoking ./manage.py update_index).
However, that change results in infinite loop index updates - one single node update will make worker spend in infinite loop updating whole index over and over.
Yet another problem popped up - after node move - indexes of all moved nodes (i.e. moved node and all its descendants) instances should be updated as well, otherwise search results will reveal nodes with outdated breadcrumbs.
Component: backend (papermerge-core) Version: 2.1.3
While removing mptt dependency papermerge.search.tasks.update_index was changed: instead of updating index for each individual instance - the whole index is updated (similarly to invoking ./manage.py update_index).
However, that change results in infinite loop index updates - one single node update will make worker spend in infinite loop updating whole index over and over.
Yet another problem popped up - after node move - indexes of all moved nodes (i.e. moved node and all its descendants) instances should be updated as well, otherwise search results will reveal nodes with outdated breadcrumbs.