Commit Graph

7 Commits

Author SHA1 Message Date
overtrue 83bf7649b1 test(e2e): drop 79 no-op serial markers from six more suites
serial_test's `#[serial]` is an in-process mutex. cargo-nextest, this repo's authoritative runner, gives every test its own process, so the mutex is never contended; cross-process serialization comes only from a `[test-groups]` entry with `max-threads = 1`. This continues #6209 and #6213.

Six more suites, none of them bound to a test group:

  object_lambda_test                  16
  special_chars_test                  14
  quota_test                          14
  archive_download_integrity_test     13
  list_objects_v2_pagination_test     12
  version_id_regression_test          10

Five of these do appear in `.config/nextest.toml`, but only in a lane's `default-filter` or a `slow-timeout` override — neither confers serialization. `inline_fast_path_cluster_test` is left alone precisely because it *is* bound to a group.

Each suite self-isolates: every test builds its own server through `RustFSTestEnvironment::new()` (UUID temp dir, allocated port), quota_test wrapping it in `QuotaTestEnv::new()` with a UUID bucket per test. No test mutates process env — quota_test passes its variables to the child server via `start_rustfs_server_with_env` — and the only literal addresses are `127.0.0.1:0`, which asks the kernel for a free port.

All 79 markers were bare `#[serial]`, no named groups, so the `use serial_test::serial;` import goes with the last marker in each of the six files and stays untouched in the 266 markers still spread across the crate.

Pure deletion, no test renamed and no behaviour changed.

Refs backlog#1846 (T1)
2026-08-19 10:07:21 +08:00
cxymds 6cce3d60bb fix(quota): reject oversized multipart completion (#5958)
* fix(quota): reject oversized multipart completion

* fix(arch): route quota test through app facade
2026-08-11 21:30:05 +08:00
cxymds 41e262cdab test(e2e): wait for authoritative quota usage (#5816) 2026-08-08 05:44:40 +08:00
Zhengchao An a9e3613cfd fix(quota): only require decoded length for actually framed aws-chunked bodies (#4968)
PUT admission since #4928 rejected any request whose Content-Encoding declares
aws-chunked but lacks x-amz-decoded-content-length with 400 UnexpectedContent.
Whether the body is actually chunk-framed is signalled by a STREAMING-*
x-amz-content-sha256, not by the declared encoding: the s3s auth layer only
de-frames streaming payloads and already requires the decoded length for them,
so a declared-only aws-chunked request (issue #1857 clients) carries an
unframed body whose wire Content-Length is the authoritative object size.
Admit it against that length; keep failing closed for genuinely framed bodies
without a decoded length, and use the decoded length for streaming payloads
even when Content-Encoding is absent.

Refs: https://github.com/rustfs/backlog/issues/1336
2026-07-17 16:23:00 +08:00
weisd 28f57b228c feat(s3): advance parity coverage (#2278) 2026-03-24 17:29:33 +08:00
yxrxy da58f8e291 fix: Allow non-admin users to read bucket quota configuration. (#1759)
Co-authored-by: houseme <housemecn@gmail.com>
Co-authored-by: heihutu <30542132+heihutu@users.noreply.github.com>
2026-02-10 11:23:21 +08:00
yxrxy 29d86036b1 feat: implement bucket quota system (#1461)
Signed-off-by: yxrxy <1532529704@qq.com>
Co-authored-by: loverustfs <hello@rustfs.com>
2026-01-12 11:42:07 +08:00