Егор Ивков authored 2 years ago and Yaroslav Dynnikov committed 2 years ago

Relates to https://git.picodata.io/picodata/picodata/tarantool-module/-/issues/22

* Added docs for all async primitives. Mostly docs are based on tokio,
  with some corrections to account for fibers.
* Changed type signatures in some - again to make more similar to tokio.
  (`oneshot::Sender::send`, `watch::Sender::send`, `oneshot::channel`)
* Renamed `oneshot::Sender::is_dropped` to `is_closed`
* Added `watch::Sender::is_closed`
* Added `subscribe` to tests
* Moved pub modules for channels and timeout into separate files

Initially we also planned to use tokio tests, but they cannot be
integrated into our codebase as they heavily rely on environment being
multithreaded and therefore write one big test for each primitive and
run it with [loom](https://github.com/tokio-rs/loom)

Provide generic `UnknownEnumVariant` struct instead of specifying
personalized `FromStr::Err` for every enum.

Also provide documentation and tests.

Alexey Protsenko authored 2 years ago and Yaroslav Dynnikov committed 2 years ago

This patch adopts changes from picodata CI.
The whole CI is refactored. The following problems are solved:

1. Couldn't test base image locally without overriding the old one.

The testing stage is run in two ways depending on whether some
of the docker-related files were changed.

If they didn't (most likely), testing stage is run on the `latest`
docker image as it used to.

Otherwise, the docker image is rebuilt locally and tested. This is done
by using `image:pull_policy` option, see [1].

It also makes a problem of using protected credentials irrelevant.

2. Gitlab caching used to be slow

Stupid gitlab caching (see [2]) wasted on (un)zipping cache 4 minutes
(of 6 total). Now it's replaced with a manual one: an archive (tar) is
placed in a docker volume on a self-hosted gitlab runner with no
compression.

4. No fine-grained time tracking was available

Gitlab only shows the total time of `step_script` without splitting it
into particular commands. This problem is solved by incorporating
collapsible log sections feature, see [3]. It makes logs fancy too.

Refs:

- [1] https://docs.gitlab.com/ee/ci/yaml/#imagepull_policy
- [2] https://gitlab.com/gitlab-org/gitlab-runner/-/issues/3523
- [3] https://docs.gitlab.com/ee/ci/jobs/#custom-collapsible-sections

Before this patch, when an attempt to lock a mutex failed, only the
previous lock location was printed. But we still had no idea where that
fail occured.

With this patch the log message looks as follows:

```
can't lock mutex at src/traft/node.rs:234:14,
already locked at src/traft/node.rs:876:44
```

Also, this patch makes the corresponding test more strict.

When lua raises an error from the code loaded with `tlua::exec()` and
`tlua::eval()`, it looks as follows:

```
[string \"chunk\"]:1: oops
```

It gives no useful information and complicates debugging.

With this patch, lua errors will correctly point to the place where
`exec()` was called, e.g.:

```
[tests/src/tlua/lua_functions.rs:216]:6: oops
```

godzie44 authored 2 years ago and Georgy Moshkin committed 2 years ago

feature(tuple): add tuple::as_named_buffer function - returns messagepack encoded tuple with named fields

It was unimplemented and will never be. The development continues as a
part of picodata product.

This is mainly useful for capturing the lua context passed to
rust-callbacks for example for use with `tlua::error!`.

See test `error` in `tests/src/tlua/functions_write.rs` for examples.

Georgy Moshkin authored 2 years ago and Georgy Moshkin committed 2 years ago

The clarification is about safety of using borrowed arguments in stored
procedures: borrowed arguments aren't if used after calls to tarantool
api.