flan
997c257442
Jellyfin treats a rating string it cannot parse the same as no rating at all: the item takes the unrated path and stays visible to every profile, including one with an age ceiling set. This library carried eighteen such items. Six were Singapore adult ratings (M18, NC16, R21) that no age ceiling would have caught, and two (TP, L) mean "all ages" and were being withheld from children for the opposite reason. The task remaps those strings to their US equivalents using published board equivalences, and locks the repaired value against provider refresh. Ratings are never inferred from genre or synopsis; an unmapped string is reported as a warning instead. A wrong guess there fails open, which is the one outcome parental controls exist to prevent. The audience tag backs a per-user Allowed Tags whitelist. That check fails closed, so new content is invisible to the filtered account until it is tagged, which is why the work has to recur rather than run once. A ledger records every id the task has tagged; an id present in the ledger but no longer carrying the tag was untagged deliberately and is never tagged again. Running in-process as a scheduled task means no API key, no credential file and no container. It also sidesteps three defects in the 12.0 REST item-update path, which assigns Name, Overview, ProductionYear, OfficialRating and CustomRating unconditionally, silently unlocks an item when LockData is omitted, and rejects its own serialized Trickplay block on the way back in.
7 lines
35 B
Plaintext
7 lines
35 B
Plaintext
bin/
|
|
obj/
|
|
*.user
|
|
.vs/
|
|
.idea/
|
|
*.zip
|