cjw/Oxicloud
1
0
Fork 0
Code Issues Pull Requests Actions 4 Packages Projects Releases Wiki Activity

Merge pull request #596 from swissiety/rfc-5789-http-patch

Browse Source
This commit is contained in:
Dionisio Pozo
2026-07-22 00:32:27 +02:00
committed by GitHub
parent 95b22464ae ec2b533a53
commit 67fe944c2a
20 changed files with 2921 additions and 133 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/interfaces/api/handlers/webdav_handler.rs
+521 -35
View File
@@ -14,7 +14,9 @@ use axum::{
};
use bytes::{Buf, Bytes};
use chrono::Utc;
use futures::stream::{self, Stream};
use quick_xml::Writer;
use std::pin::Pin;
use uuid::Uuid;
use crate::application::adapters::webdav_adapter::{
@@ -29,6 +31,7 @@ use crate::application::ports::file_ports::{FileManagementUseCase, FileUploadUse
use crate::application::ports::folder_ports::FolderUseCase;
use crate::application::ports::storage_ports::StorageUsagePort;
use crate::application::services::file_retrieval_service::FileRetrievalService;
use crate::application::services::file_upload_service::FileUploadService;
use crate::application::services::folder_service::FolderService;
use crate::common::di::AppState;
use crate::domain::repositories::drive_repository::DriveRepository;
@@ -38,6 +41,7 @@ use crate::infrastructure::services::webdav_dead_property_store::{DeadPropertySt
use crate::interfaces::errors::AppError;
use crate::interfaces::middleware::auth::{AuthUser, CurrentUser};
use crate::interfaces::range_requests::{not_modified_response, range_response};
use crate::interfaces::upload_ingest::{IngestedBlob, RangeSegment, discard_ingested};
use percent_encoding::{AsciiSet, NON_ALPHANUMERIC, percent_decode_str, utf8_percent_encode};
use std::collections::HashMap;
use std::sync::Arc;
@@ -408,6 +412,7 @@ async fn handle_webdav_dispatch(
"GET" => handle_get(state, req, path).await,
"HEAD" => handle_head(state, req, path).await,
"PUT" => handle_put(state, req, path).await,
"PATCH" => handle_patch(state, req, path).await,
"MKCOL" => handle_mkcol(state, req, path).await,
"DELETE" => handle_delete(state, req, path).await,
"MOVE" => handle_move(state, req, path).await,
@@ -439,7 +444,7 @@ async fn handle_options(_path: String) -> Result<Response<Body>, AppError> {
.header(HEADER_DAV, "1, 2") // Class 1 and 2 WebDAV support
.header(
header::ALLOW,
"OPTIONS, GET, HEAD, PUT, DELETE, PROPFIND, PROPPATCH, MKCOL, COPY, MOVE, LOCK, UNLOCK",
"OPTIONS, GET, HEAD, PUT, PATCH, DELETE, PROPFIND, PROPPATCH, MKCOL, COPY, MOVE, LOCK, UNLOCK",
)
.body(Body::empty())
.unwrap())
@@ -1655,7 +1660,7 @@ fn evaluate_if_header(
///
/// Shared by `handle_put`, `handle_delete`, `handle_move`,
/// `handle_copy`, and `handle_proppatch`.
fn enforce_native_lock(
pub(crate) fn enforce_native_lock(
lock_store: &crate::infrastructure::services::webdav_lock_service::WebDavLockStore,
if_header: Option<&str>,
path: &str,
@@ -1891,39 +1896,20 @@ async fn handle_put(
// ── RFC 7232 conditional preconditions ────────────────────────────
// Evaluated before ingesting the body to save bandwidth on doomed requests.
if let Some(ref inm) = if_none_match {
// If-None-Match: * → fail if resource exists (prevent overwrite)
if inm == "*" && file_existed {
return Err(AppError::precondition_failed(
"If-None-Match: * — resource already exists",
));
}
// Shared with `handle_patch` (both surfaces) — handles comma-separated
// multi-value lists and the weak/strong distinction the previous
// hand-rolled single-tag comparison here didn't.
if let Some(ref value) = if_none_match
&& if_none_match_precondition_fails(value, current_etag.as_deref())
{
return Err(AppError::precondition_failed(
"If-None-Match — resource already exists with that ETag",
));
}
if let Some(ref im) = if_match {
if im == "*" {
// If-Match: * → fail if resource does not exist
if !file_existed {
return Err(AppError::precondition_failed(
"If-Match: * — resource does not exist",
));
}
} else {
// If-Match: <etag> → strong comparison against current ETag
match &current_etag {
None => {
return Err(AppError::precondition_failed(
"If-Match — resource does not exist",
));
}
Some(etag) => {
let client_tag = im.trim_matches('"');
let server_tag = etag.trim_matches('"');
if client_tag != server_tag {
return Err(AppError::precondition_failed("If-Match — ETag mismatch"));
}
}
}
}
if let Some(ref value) = if_match
&& if_match_precondition_fails(value, current_etag.as_deref())
{
return Err(AppError::precondition_failed("If-Match — ETag mismatch"));
}
// ── Streaming ingest ──────────────────────────────────────────────
@@ -1971,6 +1957,7 @@ async fn handle_put(
&content_type,
None,
user.id,
None,
)
.await;
@@ -1986,7 +1973,7 @@ async fn handle_put(
};
Ok(Response::builder()
.status(status)
.header(header::ETAG, &file_dto.etag)
.header(header::ETAG, format!("\"{}\"", file_dto.etag))
.body(Body::empty())
.unwrap())
}
@@ -1998,6 +1985,412 @@ async fn handle_put(
}
}
/// Parses the `X-Update-Range` header used by [`handle_patch`] (RFC 5789
/// partial content updates): either `append`, or `bytes=<start>-<end>`
/// (inclusive, 0-based). For the explicit-range form both bounds must fall
/// strictly within the current file size — growing the file via a byte
/// range isn't supported, use `append` or PUT for that.
///
/// Returns `(start, end)`; `end` is `None` for `append`.
///
/// `pub(crate)` so the NextCloud-surface PATCH handler
/// (`nextcloud/webdav_handler.rs::handle_patch`) can reuse it instead of
/// duplicating the parsing logic.
pub(crate) fn parse_update_range(header: &str, size: u64) -> Result<(u64, Option<u64>), AppError> {
let header = header.trim();
if header.eq_ignore_ascii_case("append") {
return Ok((size, None));
}
let spec = header.strip_prefix("bytes=").ok_or_else(|| {
AppError::bad_request("X-Update-Range must be 'append' or 'bytes=<start>-<end>'")
})?;
let (start_str, end_str) = spec
.split_once('-')
.ok_or_else(|| AppError::bad_request("X-Update-Range must be 'bytes=<start>-<end>'"))?;
let start: u64 = start_str
.parse()
.map_err(|_| AppError::bad_request("X-Update-Range: invalid start offset"))?;
let end: u64 = end_str
.parse()
.map_err(|_| AppError::bad_request("X-Update-Range: invalid end offset"))?;
if start > end {
return Err(AppError::bad_request(
"X-Update-Range: start must be <= end",
));
}
if end >= size {
return Err(AppError::new(
StatusCode::RANGE_NOT_SATISFIABLE,
format!("X-Update-Range end {end} is out of bounds for a {size}-byte file"),
"RangeNotSatisfiable",
));
}
Ok((start, Some(end)))
}
/// Strip the optional `W/` weak prefix and surrounding double-quotes
/// from one ETag value in an `If-Match` / `If-None-Match` list. Returns
/// `(is_weak, inner)`.
fn parse_etag_value(raw: &str) -> (bool, &str) {
let trimmed = raw.trim();
if let Some(rest) = trimmed.strip_prefix("W/") {
(true, rest.trim().trim_matches('"'))
} else {
(false, trimmed.trim_matches('"'))
}
}
/// RFC 7232 §3.2 — `If-None-Match` fails when:
/// - the header value is `*` and a current representation exists, OR
/// - any listed ETag matches the current representation (weak comparison
/// — weak validators in the request are equivalent to strong for the
/// match itself, only If-Match is required to be strong).
///
/// `pub(crate)` so both the plain and NextCloud-surface WebDAV handlers
/// share one RFC 7232-conformant implementation instead of each
/// reimplementing ETag comparison (multi-value lists, `W/` weak prefix).
pub(crate) fn if_none_match_precondition_fails(header: &str, current_etag: Option<&str>) -> bool {
let v = header.trim();
if v == "*" {
return current_etag.is_some();
}
let Some(current) = current_etag else {
return false;
};
v.split(',').any(|tag| {
let (_, parsed) = parse_etag_value(tag);
!parsed.is_empty() && parsed == current
})
}
/// RFC 7232 §3.1 — `If-Match` fails when:
/// - the resource doesn't currently exist (no strong validator to match), OR
/// - the header isn't `*` and no listed ETag strong-matches the current one
/// (weak validators in the request never satisfy a strong-match).
pub(crate) fn if_match_precondition_fails(header: &str, current_etag: Option<&str>) -> bool {
let v = header.trim();
let Some(current) = current_etag else {
return true;
};
if v == "*" {
return false;
}
!v.split(',').any(|tag| {
let (is_weak, parsed) = parse_etag_value(tag);
!is_weak && !parsed.is_empty() && parsed == current
})
}
/// Build the untouched prefix/suffix byte-range streams either side of a
/// PATCH edit, paired with their known lengths (`upload_ingest::RangeSegment`)
/// ready to hand to `ingest_range_patch_to_cas`.
///
/// `pub(crate)` so both the plain and NextCloud-surface PATCH handlers share
/// one implementation instead of each re-deriving the same offsets — this was
/// byte-identical duplicated code before the DRY pass that added this fn.
pub(crate) async fn splice_patch_streams(
file_retrieval: &FileRetrievalService,
file_id: &str,
caller_id: Uuid,
start: u64,
end: Option<u64>,
file_size: u64,
) -> Result<(RangeSegment, RangeSegment), AppError> {
let prefix_stream: Pin<Box<dyn Stream<Item = Result<Bytes, std::io::Error>> + Send>> =
if start == 0 {
Box::pin(stream::empty())
} else {
Box::into_pin(
file_retrieval
.get_file_range_stream_with_perms(file_id, caller_id, 0, Some(start))
.await
.map_err(AppError::from)?,
)
};
let suffix_len = match end {
Some(end) if end + 1 < file_size => file_size - (end + 1),
_ => 0,
};
let suffix_stream: Pin<Box<dyn Stream<Item = Result<Bytes, std::io::Error>> + Send>> = match end
{
Some(end) if end + 1 < file_size => Box::into_pin(
file_retrieval
.get_file_range_stream_with_perms(file_id, caller_id, end + 1, None)
.await
.map_err(AppError::from)?,
),
_ => Box::pin(stream::empty()),
};
Ok(((prefix_stream, start), (suffix_stream, suffix_len)))
}
/// Quota-check + compare-and-swap write for a PATCH edit already spliced and
/// ingested into the chunk store (`ingested`). On quota rejection the blob
/// reference just taken by ingest is released and `QuotaExceeded` (507) is
/// returned; on success this is the CAS write keyed on `expected_hash` (the
/// file's pre-splice content hash) that closes the race between two
/// concurrent PATCHes to disjoint ranges of the same file (see
/// `FileBlobWritePort::swap_blob_hash`).
///
/// `pub(crate)` — shared by the plain and NextCloud-surface PATCH handlers;
/// `log_prefix` lets each surface keep its own log-line tag (`"WEBDAV PATCH"`
/// vs `"NC WEBDAV PATCH"`) for grep-ability.
#[allow(clippy::too_many_arguments)]
pub(crate) async fn cas_write_patch(
state: &AppState,
upload_service: &FileUploadService,
path: &str,
drive_id: Uuid,
ingested: &IngestedBlob,
caller_id: Uuid,
expected_hash: &str,
log_prefix: &str,
) -> Result<FileDto, AppError> {
if let Some(storage_svc) = state.storage_usage_service.as_ref()
&& let Err(err) = storage_svc
.check_storage_quota(caller_id, ingested.size)
.await
{
discard_ingested(&state.core.dedup_service, ingested).await;
tracing::warn!(
"⛔ {} REJECTED (quota): user={}, file={}, size={}",
log_prefix,
caller_id,
path,
ingested.size
);
return Err(AppError::new(
StatusCode::INSUFFICIENT_STORAGE,
err.message,
"QuotaExceeded",
));
}
let content_type = ingested.content_type.clone();
upload_service
.update_file_streaming_with_perms(
path,
drive_id,
ingested.stored(),
&content_type,
None,
caller_id,
Some(expected_hash),
)
.await
.map_err(AppError::from)
}
/**
* Handles PATCH requests (RFC 5789) for partial byte-range content updates.
*
* RFC 4918 §9.7.1 forbids partial content updates on PUT (see the explicit
* `Content-Range` rejection in [`handle_put`]); PATCH is the mechanism this
* server offers instead, via the `X-Update-Range` header (see
* [`parse_update_range`]).
*
* The new content is assembled by splicing the request body between the
* file's untouched prefix/suffix byte ranges and re-ingesting the result as
* one continuous stream through the same content-addressable pipeline PUT
* uses ([`upload_ingest::ingest_range_patch_to_cas`]) — unedited chunks on
* either side of the edit typically dedup for free.
*
* @param state The application state containing service dependencies
* @param req The HTTP request containing the partial content and
* `X-Update-Range` header
* @param path The requested resource path
* @return HTTP response: 204 with `Content-Range`/`ETag` on success
*/
async fn handle_patch(
state: Arc<AppState>,
req: Request<Body>,
path: String,
) -> Result<Response<Body>, AppError> {
use crate::interfaces::upload_ingest;
let user = extract_user(&req)?;
let file_upload_service = &state.applications.file_upload_service;
let file_retrieval_service = &state.applications.file_retrieval_service;
if path.is_empty() || path == "/" {
return Err(AppError::bad_request("Cannot PATCH the root folder"));
}
// RFC 5789 doesn't define Content-Range semantics; this server uses a
// dedicated `X-Update-Range` header instead (see `parse_update_range`)
// to avoid ambiguity with HTTP Range-Request semantics.
if req.headers().contains_key(header::CONTENT_RANGE) {
return Err(AppError::bad_request(
"PATCH must not use Content-Range; use the X-Update-Range header instead",
));
}
let update_range_header = req
.headers()
.get("X-Update-Range")
.and_then(|v| v.to_str().ok())
.map(|s| s.to_string())
.ok_or_else(|| AppError::bad_request("PATCH requires an X-Update-Range header"))?;
// Extract all headers before consuming `req` into the body stream.
let if_header_owned = req
.headers()
.get("If")
.and_then(|v| v.to_str().ok())
.map(|s| s.to_string());
let if_none_match = req
.headers()
.get(header::IF_NONE_MATCH)
.and_then(|v| v.to_str().ok())
.map(|s| s.trim().to_string());
let if_match = req
.headers()
.get(header::IF_MATCH)
.and_then(|v| v.to_str().ok())
.map(|s| s.trim().to_string());
let content_length = req
.headers()
.get(header::CONTENT_LENGTH)
.and_then(|v| v.to_str().ok())
.and_then(|s| s.parse::<u64>().ok());
let content_type = req
.headers()
.get(header::CONTENT_TYPE)
.and_then(|v| v.to_str().ok())
.unwrap_or("application/octet-stream")
.to_string();
let max_upload = state.core.config.storage.direct_put_max_bytes;
let scope = resolve_webdav_scope_or_405(&state, user.id, &path).await?;
let drive_id = scope.drive_id;
let path = scope.db_path;
// ── Existence check ───────────────────────────────────────────────
// Unlike PUT, PATCH requires an existing file — a partial update of
// nothing isn't meaningful. Resolver is drive-scoped, not
// owner-scoped (see `handle_put`'s identical comment), so the
// explicit `authz.require(Read, …)` below is the defence-in-depth
// existence-proof before any field of `file` is trusted.
let resolver = state.path_resolver.as_ref().ok_or_else(|| {
AppError::method_not_allowed("PATCH requires WebDAV path resolver support")
})?;
let file = match resolver.resolve_path_in_drive(&path, drive_id).await {
Ok(ResolvedResource::File(f)) => f,
Ok(ResolvedResource::Folder(_)) => {
return Err(AppError::conflict("Cannot PATCH a directory"));
}
Err(_) => return Err(AppError::not_found(format!("File not found: {}", path))),
};
let file_uuid = Uuid::parse_str(&file.id)
.map_err(|_| AppError::not_found(format!("File not found: {}", path)))?;
state
.authorization
.require(
Subject::User(user.id),
Permission::Read,
Resource::File(file_uuid),
)
.await?;
// ── Active-lock guard + RFC 4918 §10.4 If: evaluation ─────────────
if let Some(resp) = enforce_native_lock(
&state.webdav_lock_store,
if_header_owned.as_deref(),
&path,
Some(&file.etag),
) {
return Ok(resp);
}
// ── RFC 7232 conditional preconditions ────────────────────────────
let current_etag = Some(file.etag.as_str());
if let Some(ref value) = if_none_match
&& if_none_match_precondition_fails(value, current_etag)
{
return Err(AppError::precondition_failed(
"If-None-Match — resource already exists with that ETag",
));
}
if let Some(ref value) = if_match
&& if_match_precondition_fails(value, current_etag)
{
return Err(AppError::precondition_failed("If-Match — ETag mismatch"));
}
// ── Range parsing + validation ─────────────────────────────────────
let (start, end) = parse_update_range(&update_range_header, file.size)?;
if let (Some(end), Some(len)) = (end, content_length) {
let expected = end - start + 1;
if len != expected {
return Err(AppError::bad_request(format!(
"Content-Length {len} does not match X-Update-Range span {expected}"
)));
}
}
// ── Splice prefix/suffix around the patched span ───────────────────
let (prefix_segment, suffix_segment) = splice_patch_streams(
file_retrieval_service,
&file.id,
user.id,
start,
end,
file.size,
)
.await?;
let filename = crate::common::mime_detect::filename_from_path(&path).to_string();
let ingested = upload_ingest::ingest_range_patch_to_cas(
prefix_segment,
req.into_body(),
suffix_segment,
&state.core.dedup_service,
&filename,
&content_type,
upload_ingest::PatchIngestBudget {
max_bytes: max_upload,
expected_body_len: end.map(|end| end - start + 1),
},
)
.await?;
// ── Quota enforcement + atomic store, compare-and-swap on the
// pre-splice content hash ─────────────────────────────────────────
// `file.content_hash` was snapshotted before the (potentially slow)
// splice + CAS-ingest above. Passing it as `expected_hash` makes the
// write itself a compare-and-swap: the repository checks and applies
// under the same row lock, so nothing else can write to this file
// between the check and the write. This is what actually closes the
// race two concurrent PATCHes to disjoint ranges could otherwise hit
// — each individually passing its own If-Match check against the
// same stale snapshot, then blindly overwriting each other.
let new_size = ingested.size;
let file_dto = cas_write_patch(
&state,
file_upload_service,
&path,
drive_id,
&ingested,
user.id,
&file.content_hash,
"WEBDAV PATCH",
)
.await?;
// Everything from `start` to the new EOF reflects the patch
// (the untouched suffix, if any, may have shifted when the
// body's length differs from the replaced span).
let range_end = new_size.saturating_sub(1);
Ok(Response::builder()
.status(StatusCode::NO_CONTENT)
.header(header::ETAG, format!("\"{}\"", file_dto.etag))
.header(
header::CONTENT_RANGE,
format!("bytes {}-{}/{}", start, range_end, new_size),
)
.body(Body::empty())
.unwrap())
}
/**
* Handles MKCOL requests to create folders.
*
@@ -3385,4 +3778,97 @@ mod tests {
"/webdav/My%20Photos/2024/"
);
}
// ── RFC 5789 PATCH: `X-Update-Range` parsing ────────────────────
#[test]
fn parse_update_range_append() {
assert_eq!(parse_update_range("append", 100).unwrap(), (100, None));
assert_eq!(parse_update_range("APPEND", 0).unwrap(), (0, None));
}
#[test]
fn parse_update_range_explicit_span() {
assert_eq!(parse_update_range("bytes=5-9", 100).unwrap(), (5, Some(9)));
// Single-byte span at offset 0.
assert_eq!(parse_update_range("bytes=0-0", 1).unwrap(), (0, Some(0)));
}
#[test]
fn parse_update_range_rejects_missing_prefix() {
assert!(parse_update_range("5-9", 100).is_err());
}
#[test]
fn parse_update_range_rejects_malformed_bounds() {
assert!(parse_update_range("bytes=abc-9", 100).is_err());
assert!(parse_update_range("bytes=5-abc", 100).is_err());
assert!(parse_update_range("bytes=9", 100).is_err());
}
#[test]
fn parse_update_range_rejects_start_after_end() {
assert!(parse_update_range("bytes=9-5", 100).is_err());
}
#[test]
fn parse_update_range_rejects_end_at_or_past_size() {
// `end` must be strictly within the current file — growing the
// file via a byte-range PATCH isn't supported (use `append`).
let err = parse_update_range("bytes=5-9", 9).unwrap_err();
assert_eq!(err.status_code, StatusCode::RANGE_NOT_SATISFIABLE);
assert!(parse_update_range("bytes=0-0", 0).is_err());
}
// ── RFC 7232 If-Match / If-None-Match — multi-value lists ───────
//
// Regression coverage for `handle_put`'s hand-rolled precondition
// check, which only ever compared the header as a single tag and
// never split on commas — a client sending the standard
// comma-separated multi-value form would silently mismatch even
// when one of the listed ETags matched. Both handlers now share
// `if_none_match_precondition_fails`/`if_match_precondition_fails`,
// which already handled this correctly for `handle_patch`.
#[test]
fn if_none_match_multi_value_list_matches_second_tag() {
assert!(if_none_match_precondition_fails(
r#""aaa", "bbb", "ccc""#,
Some("bbb")
));
}
#[test]
fn if_none_match_multi_value_list_no_match_passes() {
assert!(!if_none_match_precondition_fails(
r#""aaa", "bbb", "ccc""#,
Some("zzz")
));
}
#[test]
fn if_match_multi_value_list_matches_last_tag() {
assert!(!if_match_precondition_fails(
r#""aaa", "bbb", "ccc""#,
Some("ccc")
));
}
#[test]
fn if_match_multi_value_list_no_match_fails() {
assert!(if_match_precondition_fails(
r#""aaa", "bbb", "ccc""#,
Some("zzz")
));
}
#[test]
fn if_match_weak_tag_in_list_never_satisfies() {
// If-Match requires a strong comparison — a weak validator in the
// list must not satisfy it even if the underlying tag matches.
assert!(if_match_precondition_fails(
r#"W/"aaa", "bbb""#,
Some("aaa")
));
}
}
src/interfaces/api/handlers/wopi_handler.rs
+1
View File
@@ -413,6 +413,7 @@ async fn put_file(
&content_type,
None,
claims_sub_uuid,
None,
)
.await;