- As a student, upload four documents to an assignment:
- Shell script
All valid documents are converted
No documents are converted
This isn't as easy as it looks. The problem comes if a regeneration is forced for a document when the supported document list has changed.
We don't know the number of pages in that shell script, and if it happens in the middle... well, it means that we have to be clever about shifting existing annotations.
Annotations are currently against the whole converted document. It is unaware of the constituent parts, of how many pages they have, etc. If we insert a new page in the middle, or remove an existing one, we currently do not move the annotations to keep with the relevant PDF.
Basically, we'd need to either make the annotations document-aware, or make the document_services API more aware of the annotations and page counts at the point that we convert a document.