GetProfileItemsOrderedByIndexIds calls Distinct() on the incoming indexIds, which removes duplicate ids. Previously, call sites that built ordered lists (e.g., group ChildItems ordering) would preserve duplicates by iterating the original id list. If duplicates are meaningful (e.g., weighting / repeated nodes), this changes behavior; consider keeping the original idList for ordering and using a separate distinct set only for the DB query.

这边就不去重了？

去重应该更合理些

但是是否存在一个可能，用户是故意添加的重复的呢？

那就不改了

并且这玩意也不是严格对应的，会跳过某些无效项，也不应该直接下标访问，注释没写

但是是否存在一个可能，用户是故意添加的重复的呢？

先不管了吧，重复的没什么意义，有人说了再改吧

The method signature takes IEnumerable<string> indexIds, but several new call sites pass Select(sp => sp?.IndexId) which is naturally an IEnumerable<string?> and can contain nulls. Since the implementation already filters out null/empty ids, consider changing the parameter type to IEnumerable<string?> (or accepting nullable and guarding) to better reflect actual usage and avoid nullability mismatches.

idList.Select(id => itemMap.GetValueOrDefault(id)).Where(item => item != null).ToList() relies on a null-filter to satisfy the List<ProfileItem> return type. To keep the non-null contract explicit (and resilient if nullable analysis is enabled later), consider using OfType<ProfileItem>() or null-forgiving after the filter so the resulting list is statically non-nullable.

这个刚开始的写法就是 .OfType<ProfileItem>() 来着，ReSharper 直接冒绿光让我改成 .Where(item => item != null)

按编辑器的提示吧。 ai 有时候会弄过时的方案