A small release: automatic metadata fetch now fills in a missing series from another source, books with non-ASCII filenames upload again, and long series names wrap on book cards in the new UI.
docker pull ghcr.io/new-usemame/calibre-web-nextgen:v4.1.53Before you upgrade
- Nothing to do. There are no database changes, and the KOReader plugin stays at 4.1.48.
Fixed
- Automatic metadata fetch fills in a missing series from another source. When the source that matched a book gives no series (Google Books, Open Library and IBDb never do), up to three of the series sources in your provider order are asked next, then Hardcover and Kobo if your order leaves them out. Only the series and its position are taken, and only from a result that is the same book. Page scrapers such as Amazon and Goodreads are asked only when you put them in the order. Thanks @splitsec2 (#2684).
- Another source's spelling of a series you already have joins it. "Dune Chronicles" or "Red Rising Saga" now land in the library's "Dune" and "Red Rising" instead of starting a second series. Names with two different notes, such as "Star Wars (Legends)" and "Star Wars (Canon)", stay separate.
- Series positions such as 24.5 are no longer rounded down from Kobo and Amazon, and Goodreads books that only name the series in the title, as in "Cross Kill (Alex Cross, #23.5)", now get that series.
- Uploading a book whose filename is all non-ASCII no longer loses it. A file such as
龙蛋.epubwas staged without its extension, so the import step deleted it after the upload had been reported as queued. CJK, Cyrillic and other non-ASCII filenames now import normally. Reported by @HernandoR in #2701. - A long series name under a book in the new UI wraps to a second line instead of ending in an ellipsis, and the series number stays visible. Cards in a row stay aligned. Reported by @mrfearless in #2147, fixed by @sgreadly (#2697).
Full details in CHANGELOG.md.