Chunking + Map-first preprocessing + branch filtering #chunk #branch-filtering
We notice that ditamaps with a ditavalref element at the root of their map and chunks at some chapter maps end up with broken links in normalized output for those chapter maps. It looks like the chunking adds the root element ID at the end of the hashed filenames in the topicref’s href attribute and one of the preprocessing steps fails to resolve the hash reference back to the original file name.
Has anyone seen this? Also, does anyone know where that logic exists in the preprocess2 pipeline? Ideally we could override it to simply use the hashed filename without the appended ID value or override the chunk preprocessing not to add it at all.
Sent from mobile