Commit graph

11 commits

Author SHA1 Message Date
a.r
d9969a9f4f
lazy_static → once_cell → stabilized versions (#3447)
* Anki: Replace lazy_static with once_cell

Unify to once_cell, lazy_static's replacement. The latter in unmaintained.

* Anki: Replace once_cell with stabilized LazyCell / LazyLock as far as possible

Since 1.80: https://github.com/rust-lang/rust/issues/109736 and https://github.com/rust-lang/rust/pull/98165

Non-Thread-Safe Lazy → std::cell::LazyCell https://doc.rust-lang.org/nightly/std/cell/struct.LazyCell.html

Thread-safe SyncLazy → std::sync::LazyLock https://doc.rust-lang.org/nightly/std/sync/struct.LazyLock.html

The compiler accepted LazyCell only in minilints.

The final use in rslib/src/log.rs couldn't be replaced since get_or_try_init has not yet been standardized: https://github.com/rust-lang/rust/issues/109737

* Declare correct MSRV (dae)

Some of our deps require newer Rust versions, so this was misleading.

Updating the MSRV also allows us to use .inspect() on Option now
2024-09-30 23:35:06 +10:00
Damien Elmes
5335d748cf Bump Rust to 1.81 for latest CVE 2024-09-05 12:39:59 +07:00
Jean Khawand
83fe301c1c
Add distroless Dockerfile and implement internal health check (#3366)
- rslib(http_server): add `is_running()` method
- rslib(sync): introduce `--healthcheck` argument for health probe in distroless
- doc(syncserver): add table comparing Dockerfile and Dockerfile.distroless
- Expand cross-platform support with distroless
- add `Dockerfile.distroless`

- Dockerfile: bump rust `1.79` to `1.80.1`
- Dockerfile: bump alpine `3.20` to `3.20.2`

Note: Implemented an internal health check because distroless images do not include curl, which is used to reduce image size and attack surface. For more details, see https://blog.sixeyed.com/docker-healthchecks-why-not-to-use-curl-or-iwr/
https://github.com/GoogleContainerTools/distroless

fix: failed: check:format:rust

typo

remove extra space

fix failed:check:format:rust

update doc

fetch `host` and `port` using envy

fix: failed: check:format:rust

Update doc + add dockerignore

- dockerignore: This helps avoid sending unwanted files and directories to the builder
- add new line
- I am still experimenting cross platform compilation, I am getting
4.337 From https://github.com/ankitects/rust-url
4.337  * [new ref]         bb930b8d089f4d30d7d19c12e54e66191de47b88 -> refs/commit/bb930b8d089f4d30d7d19c12e54e66191de47b88
4.397 error: failed to get `percent-encoding-iri` as a dependency of package `anki v0.0.0 (/app/rslib)`

still checking what could be the issue

fix: failed: check:format:dprint
2024-08-29 17:05:33 +07:00
Jean Khawand
56a085bc21
Update base images and introduce health endpoint (#3273)
* Update base images and introduce health endpoint

sync-server: introduce `/health` endpoint to check if the service is reachable.

bump(alpine): bump alpine base image from `3.19` to `3.20`

bump(rust): bump rust-alpine build image from `1.76` to `1.79`

* fix cargo fmt

* add allow clippy::extra_unused_type_parameters

* Remove unused type param (dae)

* Route /health directly (dae)

* Fix for latest axum (dae)
2024-07-10 20:35:21 +07:00
Damien Elmes
227f790d8e Update axum/reqwest for hyper 1.0 2024-07-06 18:45:32 +07:00
laalsaas
798d9df4de
sync-server: add option to store hashed passwords (#3083) 2024-03-26 07:39:57 +00:00
Damien Elmes
bfef908c6c
Refactor media sync handling (#2647)
* Refactor media sync handling

- The media USN is now returned in sync/meta, which avoids an extra
round-trip.
- Media syncing is now automatically started by the syncing code at
the end of a normal or full sync, which avoids it competing for bandwidth
and resources, and avoids duplicate invalid login messages when the auth
token is invalid.
- Added a new media_sync_progress() method to both check if media is
syncing, and get access to the latest progress.
- Updated the sync log screen to only show the latest line, like AnkiMobile.
- Show media sync errors in a pop-up, so they don't get missed. Use a non-modal
pop-up to avoid potential conflicts with other modals.

* Remove print statement
2023-09-10 13:22:20 +10:00
Damien Elmes
85c2769f80
Update Rust and Python deps (#2567)
* Update Python deps

* Update semver-compat Rust deps

* Update most crates to latest semver

* Update to latest axum-client-ip
2023-07-01 18:26:43 +10:00
Damien Elmes
d380f3034c Split io.rs into separate crate, and use it in proto build
Will be handy to use it in our other scripts in the future too - thanks
Rumo!

Results of benchmarking ./run before and after these crate splits:

- Touching a proto file leads to a slight increase: about +90ms
- Touching an rslib file leads to a bigger decrease, as there's less to
recompile: about -700ms

And ./ninja test is even better: about +200ms and -3800ms.
2023-06-12 15:47:51 +10:00
Damien Elmes
ded805b504
Switch Rust import style (#2330)
* Prepare to switch Rust import style

* Run nightly format

Closes #2320

* Clean up a few imports

* Enable comment wrapping

* Wrap comments
2023-01-18 21:39:55 +10:00
Damien Elmes
cf45cbf429
Rework syncing code, and replace local sync server (#2329)
This PR replaces the existing Python-driven sync server with a new one in Rust.
The new server supports both collection and media syncing, and is compatible
with both the new protocol mentioned below, and older clients. A setting has
been added to the preferences screen to point Anki to a local server, and a
similar setting is likely to come to AnkiMobile soon.

Documentation is available here: <https://docs.ankiweb.net/sync-server.html>

In addition to the new server and refactoring, this PR also makes changes to the
sync protocol. The existing sync protocol places payloads and metadata inside a
multipart POST body, which causes a few headaches:

- Legacy clients build the request in a non-deterministic order, meaning the
entire request needs to be scanned to extract the metadata.
- Reqwest's multipart API directly writes the multipart body, without exposing
the resulting stream to us, making it harder to track the progress of the
transfer. We've been relying on a patched version of reqwest for timeouts,
which is a pain to keep up to date.

To address these issues, the metadata is now sent in a HTTP header, with the
data payload sent directly in the body. Instead of the slower gzip, we now
use zstd. The old timeout handling code has been replaced with a new implementation
that wraps the request and response body streams to track progress, allowing us
to drop the git dependencies for reqwest, hyper-timeout and tokio-io-timeout.

The main other change to the protocol is that one-way syncs no longer need to
downgrade the collection to schema 11 prior to sending.
2023-01-18 12:43:46 +10:00