Files
PiCloud/crates/shared/src/group_files.rs
MechaCat02 966b209d2e feat(quota): add per-app storage ceilings
An app's OWN store had no ceiling of any kind — only the group-SHARED
collections did, which is the inversion of what you'd expect. And
`authz::script_gate` returns `Ok(())` when there is no principal, so a
PUBLIC UNAUTHENTICATED route could reach `files::collection(..).create(..)`
and write blobs in a loop until the disk filled. `PICLOUD_FILES_MAX_FILE_SIZE_BYTES`
caps ONE file at 100 MB; nothing capped the count. Same for KV keys and docs
against Postgres.

Ceilings: `PICLOUD_APP_KV_MAX_ROWS` / `PICLOUD_APP_DOCS_MAX_ROWS` (100k) and
`PICLOUD_APP_FILES_MAX_TOTAL_BYTES` (10 GiB).

They are enforced DIFFERENTLY from the group ones, on purpose — porting the
group design as-is would have been a serious regression:

  * KV/docs check a ROW count, on INSERT only, with NO lock. `kv::set` is the
    hottest write path in the system; taking a per-app advisory lock on every
    set would serialize an app's entire data plane, and a `SUM(...)` scan would
    make write cost grow with the app's stored size. What the lock buys is
    small — unlocked overshoot is bounded by write concurrency (32) against a
    ceiling of 100k, ~0.03%. These are anti-DoS rails, not billing. An update
    adds no row and is bounded by the per-value cap, so it pays nothing at all.
  * No per-app BYTE ceiling for KV/docs: every value is already capped at
    `PICLOUD_KV_MAX_VALUE_BYTES`, so `max_rows x max_value_bytes` is ALREADY a
    finite bound. A second scan would buy nothing.
  * FILES are the exception and DO lock + sum: one blob may be 100 MB, so 32
    racing uploads could overshoot by gigabytes of real disk, and a file write
    is heavy enough that the lock and the SUM are lost in the noise. Checked on
    create AND update — an update that skipped the ceiling would be a free
    bypass (the same bug just fixed for group files: grow a 1-byte file to
    100 MB, repeat).

`group_quota` is renamed `quota`: it now serves both owners, and the module
doc is where the two enforcement strategies are contrasted.

Pinned by tests/atomic_write.rs: the key ceiling refuses a new key while
still allowing an update at the ceiling (rejecting updates would brick an app
the moment it filled up — worse than the DoS the rail exists to stop), and 10
concurrent uploads cannot exceed the disk ceiling, nor can an update grow past it.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-14 20:32:32 +02:00

210 lines
6.4 KiB
Rust

//! `GroupFilesService` — the §11.6 shared group-collection FILES contract.
//!
//! The cross-app-sharing counterpart to [`crate::FilesService`], analogous to
//! [`crate::GroupKvService`] / [`crate::GroupDocsService`]. A script reaches it
//! via the explicit `files::shared_collection("name")` handle. The owner of the
//! data is not `cx.app_id`: the impl resolves the **owning group** by walking the
//! calling app's ancestor chain for the nearest group that declares the
//! collection shared with `kind='files'`. That ancestry walk is the isolation
//! boundary — the trait surface takes **no** group id (owner derivation stays in
//! the impl).
//!
//! Trust model (§11.6): **reads are open** to any subtree script (anonymous
//! public HTTP included); **writes require an authenticated principal** with
//! editor+ on the owning group. Same shape as group KV/docs.
use async_trait::async_trait;
use thiserror::Error;
use uuid::Uuid;
use crate::{FileMeta, FileUpdate, FilesListPage, NewFile, SdkCallCx};
#[async_trait]
pub trait GroupFilesService: Send + Sync {
async fn create(
&self,
cx: &SdkCallCx,
collection: &str,
new: NewFile,
) -> Result<Uuid, GroupFilesError>;
async fn head(
&self,
cx: &SdkCallCx,
collection: &str,
id: &str,
) -> Result<Option<FileMeta>, GroupFilesError>;
async fn get(
&self,
cx: &SdkCallCx,
collection: &str,
id: &str,
) -> Result<Option<Vec<u8>>, GroupFilesError>;
async fn update(
&self,
cx: &SdkCallCx,
collection: &str,
id: &str,
upd: FileUpdate,
) -> Result<(), GroupFilesError>;
async fn delete(
&self,
cx: &SdkCallCx,
collection: &str,
id: &str,
) -> Result<bool, GroupFilesError>;
async fn list(
&self,
cx: &SdkCallCx,
collection: &str,
cursor: Option<&str>,
limit: u32,
) -> Result<FilesListPage, GroupFilesError>;
}
/// Stub for test bundles that don't exercise shared files collections.
#[derive(Debug, Default, Clone, Copy)]
pub struct NoopGroupFilesService;
#[async_trait]
impl GroupFilesService for NoopGroupFilesService {
async fn create(
&self,
_cx: &SdkCallCx,
_collection: &str,
_new: NewFile,
) -> Result<Uuid, GroupFilesError> {
Err(GroupFilesError::Backend(
"group files is not wired in".into(),
))
}
async fn head(
&self,
_cx: &SdkCallCx,
_collection: &str,
_id: &str,
) -> Result<Option<FileMeta>, GroupFilesError> {
Err(GroupFilesError::Backend(
"group files is not wired in".into(),
))
}
async fn get(
&self,
_cx: &SdkCallCx,
_collection: &str,
_id: &str,
) -> Result<Option<Vec<u8>>, GroupFilesError> {
Err(GroupFilesError::Backend(
"group files is not wired in".into(),
))
}
async fn update(
&self,
_cx: &SdkCallCx,
_collection: &str,
_id: &str,
_upd: FileUpdate,
) -> Result<(), GroupFilesError> {
Err(GroupFilesError::Backend(
"group files is not wired in".into(),
))
}
async fn delete(
&self,
_cx: &SdkCallCx,
_collection: &str,
_id: &str,
) -> Result<bool, GroupFilesError> {
Err(GroupFilesError::Backend(
"group files is not wired in".into(),
))
}
async fn list(
&self,
_cx: &SdkCallCx,
_collection: &str,
_cursor: Option<&str>,
_limit: u32,
) -> Result<FilesListPage, GroupFilesError> {
Err(GroupFilesError::Backend(
"group files is not wired in".into(),
))
}
}
/// Failure modes surfaced to the Rhai bridge. Mirrors [`crate::FilesError`] plus
/// the §11.6 `CollectionNotShared` boundary.
#[derive(Debug, Error)]
pub enum GroupFilesError {
#[error("invalid collection name: {0}")]
InvalidCollection(String),
/// No group on the calling app's ancestor chain declares this collection
/// shared with `kind='files'` — the structural "not shared with you"
/// boundary (also the outcome for a foreign app).
#[error("collection {0:?} is not a shared group files collection for this app")]
CollectionNotShared(String),
#[error("missing required field: {0}")]
MissingField(&'static str),
#[error("file too large: {size} bytes exceeds limit of {limit} bytes")]
TooLarge { size: usize, limit: usize },
/// The owning group's shared-files store is at its total-bytes quota
/// (`PICLOUD_GROUP_FILES_MAX_TOTAL_BYTES`); the write is rejected.
#[error("group files: quota exceeded ({actual} bytes; limit {limit})")]
QuotaExceeded { limit: u64, actual: u64 },
#[error("file name too long: {0} bytes exceeds 255")]
NameTooLong(usize),
#[error("content_type too long: {0} bytes exceeds 127")]
ContentTypeTooLong(usize),
#[error("file not found")]
NotFound,
#[error("file content corrupted (checksum mismatch)")]
Corrupted,
/// For reads this is only raised when `cx.principal.is_some()`; for WRITES
/// it is also raised when the principal is absent (writes fail closed).
#[error("forbidden")]
Forbidden,
#[error("group files backend error: {0}")]
Backend(String),
}
/// Map the shared [`crate::FilesError`] — returned by `validate_collection`
/// (re-exported as `validate_files_collection`) and the `NewFile`/`FileUpdate`
/// validators that the group-files service reuses — onto the group error. Lives
/// here (not in manager-core) so the orphan rule is satisfied.
impl From<crate::FilesError> for GroupFilesError {
fn from(e: crate::FilesError) -> Self {
use crate::FilesError as F;
match e {
F::InvalidCollection(c) => Self::InvalidCollection(c),
F::MissingField(f) => Self::MissingField(f),
F::TooLarge { size, limit } => Self::TooLarge { size, limit },
F::NameTooLong(n) => Self::NameTooLong(n),
F::ContentTypeTooLong(n) => Self::ContentTypeTooLong(n),
F::NotFound => Self::NotFound,
F::Corrupted => Self::Corrupted,
F::Forbidden => Self::Forbidden,
F::QuotaExceeded { limit, actual } => Self::QuotaExceeded { limit, actual },
F::Backend(s) => Self::Backend(s),
}
}
}