Let a storage backend refuse or defer a deletion, with a reason shown to the user
维护者通常 1 天内回复
还没有人认领这个 Issue。
评估
- 难度
- 5/5
- 预计耗时
- 一周以上
- 新手友好度
- 25/100
调研方向
The proposal identifies AmazonS3::unlink() and S3ObjectTrait::deleteObject() in Nextcloud storage, along with WebDAV deletion and trash emptying, but does not name file paths or tests. Start by tracing those deletion entry points and the current refusal handling; the scope spans a typed refusal, user-visible messages, deletion events, and S3 Object Lock behavior. Done means refused or deferred deletes preserve the item and explain why, while accepted deletes and S3 retention or legal holds are handled and covered by tests.
由索引模型根据 Issue 内容生成。
描述
[!TIP]
Help move this idea forward
- Use the 👍 reaction to show support for this feature.
- Avoid commenting unless you have relevant information to add; unnecessary comments create noise for subscribers.
- Subscribe to receive notifications about status changes and new comments.
Is your feature request related to a problem? Please describe.
Public bodies must keep records for their full statutory retention period, and may only destroy them through a recorded decision. A record under legal hold may not be destroyed at all. The natural setup for this is Nextcloud on top of storage that enforces retention, such as an S3 bucket with Object Lock. Today the two cannot talk to each other:
- Deletion is an unconditional instruction. A backend has no way to answer "not yet, and here is why".
- On S3 it is worse than an error.
AmazonS3::unlink()(files_external) andS3ObjectTrait::deleteObject()(primary object storage) sendDeleteObjectwithout aVersionId. Object Lock requires versioning, so S3 answers with a delete marker and reports success. The file disappears for the user and from Nextcloud, while the locked bytes stay in the bucket out of sight. Folder deletes usedeleteObjectswithQuiet: trueand do not read back the per-key errors. - When a deletion is refused today (a listener throws
AbortedEventException, or the storage returnsfalse), WebDAV answers a bare 403 and the reason is lost.
We found this by reading the code on master; we are confirming it on MinIO with Object Lock and will add the results here.
Describe the solution you'd like
- A typed refusal. A storage or listener can throw a refusal carrying a reason for the user and, optionally, the date until which the object is held. WebDAV returns it with the message, and Files shows it, also when emptying the trash.
- Permanent deletion as a request. Emptying the trash asks the backend, and the backend may refuse or defer. A refused item stays where it is.
- Events for both a refused and a completed deletion, carrying the file's identity, so audit logging can record both.
- S3 storages that respect Object Lock: recognise retention and legal hold and report them as a refusal instead of writing a delete marker, and check a folder before deleting it so it is never left half deleted.
Core keeps no retention rules. Which object is kept for how long belongs to the backend or to apps; core only understands the answer and shows it.
Describe alternatives you've considered
- files_retention. Despite the name it solves the opposite half: it deletes tagged files after a period, and nothing stops a user from deleting them earlier. Records need protection until the end date first, and only then destruction with a recorded decision. files_retention would benefit from this proposal: its background job calls
Node::delete(), so it would get a clear "not yet" back instead of silently creating a delete marker on S3. - App-level protection, such as a listener on
BeforeNodeDeletedEvent. It does not hold for records: an admin can disable the app, other code paths bypass it, and it cannot protect against direct access to the bucket. Only the storage layer can enforce retention; core needs to understand its answer.
Additional context
This is one of a small set of storage interface extensions for records-managed backends, together with backend metadata (#66355) and durable object identity (#66499). We (Conduction, working with Dutch municipalities and VNG Common Ground) would like to build and maintain the pull requests, with tests and a MinIO Object Lock test setup.
- 主要语言
- PHP
- 星标
- 37.5k
- 派生
- 5.4k
- 平均合并
- 1 天 13 小时
- 30 天内合并 PR
- 639
环境准备
在浏览器里用你自己的 GitHub 账号启动这个项目的开发容器。
- 没有 Dockerfile 或 Docker Compose 文件
- 有 Pull Request 模板
- 阅读贡献指南
从这里开始
- 先读完整个 Issue,再读项目的贡献指南。
- 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
- Fork 仓库,在一个分支上完成修改。
- 提交 Pull Request,并在描述里引用这个 Issue 编号。
nextcloud/server 的其他 Issue
-
35-feedback
难度 2/5 1-3 小时 新手友好度 78/100
维护者通常 1 天内回复
-
0. Needs triage 35-feedback bug
难度 2/5 1-3 小时 新手友好度 66/100
维护者通常 1 天内回复
-
workflowengine: File entity reuses the node of the first "Tag assigned" event for all later events in the same process可能已有人在做 @lazerg 于 2 天前认领。 未关闭
难度 2/5 1-3 小时 新手友好度 62/100
维护者通常 1 天内回复
-
Contacts import aborts with TypeError when vCards have no UID (TextImporter structure)可能已有人在做 @SebastianKrupinski 于 6 天前认领。 未关闭bug
难度 2/5 1-3 小时 新手友好度 72/100
nextcloud/server#66486 · 2 条评论 · 2 个 reaction · 已指派 1 人 ·
维护者通常 1 天内回复
-
Unified search: screen reader hears "No matching results" on mobile even when there are results可能已有人在做 @lazerg 于 10 天前认领。 未关闭
难度 2/5 1-3 小时 新手友好度 78/100
维护者通常 1 天内回复
相似的 Issue
-
[Sync EN] stream: document the errors of the stream_context_*() and stream filter functions (#6091)未关闭sync-en
难度 2/5 1-3 小时 新手友好度 66/100
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 72/100
-
难度 2/5 1-3 小时 新手友好度 79/100
Automattic/blocks-engine#6318 ·
维护者通常 1 天内回复
-
Update: Laravel Cheat Sheet - middleware examples for the bootstrap/app.php layout (small update)未关闭
难度 2/5 1-3 小时 新手友好度 84/100
OWASP/CheatSheetSeries#2990 ·
维护者通常 1 天内回复
-
难度 1/5 1-3 小时 新手友好度 82/100
维护者通常 1 天内回复