TWIR Deploy Notify

This repository contains a small tool and workflow for sending summaries of the latest This Week in Rust post to Telegram.

The GitHub Actions workflow checks out the rust-lang/this-week-in-rust repository and detects the newest Markdown file in its content directory. If a new issue is found, it is parsed with the Rust application in src/main.rs, and the generated message is posted to the configured Telegram chat. Each section becomes an individual Telegram post, and sections or overly long lines exceeding Telegram's size limit are automatically split. The parser now derives the HTML link from the issue number and date and appends it at the end of each Telegram message. A single GitHub release tagged latest always provides the most recent binary.

Toolchain

All GitHub Actions workflows read the pinned channel from rust-toolchain.toml and install the required clippy and rustfmt components automatically. The scheduled TWIR Update Rust toolchain workflow resolves the latest stable release, updates both the manifest and DOCS/RUST_VERSION.md, and opens an auto-merged pull request whenever the version changes. This keeps the repository active while ensuring every pipeline runs against the same compiler version. If the components are not installed automatically, run:

rustup component add clippy rustfmt

To run the workflow locally you must clone the this-week-in-rust repository into a twir subdirectory:

git clone --depth 1 --filter=blob:none --sparse \
  https://github.com/rust-lang/this-week-in-rust twir
cd twir && git sparse-checkout set content
cd ..

After that you can run the tool manually with:

cargo run --bin twir-deploy-notify -- twir/content/<file-name>.md

Pass --plain to generate plain text output. Tables are now rendered without code fences or extra padding by default.

All files matching output_*.md in the current directory are removed before the new posts are written.

Set RUST_LOG=info to see detailed logs including Telegram API responses:

RUST_LOG=info cargo run --bin twir-deploy-notify -- twir/content/<file-name>.md

Cargo profiles

The project defines custom profiles in Cargo.toml that optimize for quick iteration. Development and test builds use opt-level = 0 with incremental compilation enabled. Release builds set opt-level = 1, keep incremental compilation, and increase codegen-units to 16.

The .cargo/config.toml file adds -C target-cpu=native to all builds so local compilations target the current machine's CPU features.

Configuration

The application expects several environment variables when sending posts to Telegram:

• TELEGRAM_BOT_TOKEN – bot token for the selected environment.
• TELEGRAM_CHAT_ID – identifier of the chat or channel. Numeric IDs are automatically prefixed with -100 when sending requests to Telegram.
• DEV_BOT_TOKEN and DEV_CHAT_ID – credentials for the developer chat. The CLI falls back to TELEGRAM_BOT_TOKEN and TELEGRAM_CHAT_ID when the dedicated developer variables are not present, which skips the production delivery.
• TWIR_SKIP_DEVELOPER_SEND – optional boolean flag (true, false, 1, 0, etc.) that skips the developer send. Used by the production workflow because it already dispatched a developer run in a separate job.
• TWIR_SKIP_PRODUCTION_SEND – optional boolean flag that skips the production send. The developer workflow enables it to avoid double-posting when only the developer chat should receive messages.

If neither pair is fully configured, the CLI terminates with an error. Partial configuration (setting only one variable from a pair) is also treated as an error and stops the run with an actionable log message.

The first sent message is automatically pinned, and the service notification is removed. Production runs store the last processed file in last_sent.txt as an artifact and download it on the next run. The workflow now fails if this artifact cannot be retrieved.

Telegram acknowledgements are validated directly by the CLI, and the prod.yml workflow waits for the developer delivery job to succeed before the production run starts. The workflow runs hourly on the zeroth minute and publishes the latest post directly to the main chat. The retro.yml workflow builds posts for the last ten issues and uploads them as artifacts. All posts are parsed at runtime using the shared parser and generator.

Development

Continuous integration runs cargo machete to verify that Cargo.toml lists only used dependencies. Run this command locally before opening a pull request. Install it with cargo install cargo-machete if it is not available.

Documentation in DOCS/ is validated with cargo run --bin check-docs, which parses files using pulldown-cmark. Generated Telegram posts are verified with the shared validator module. Security checks using cargo-audit can be enabled in the same way by setting the run_audit input.

Auto merge

Pull requests targeting the main branch are merged automatically once all CI checks have completed. The automerge.yml workflow waits for every check run on the pull request commit to finish and fails only if a run reports a non-success conclusion. Checks that are skipped or marked as neutral do not block merging, so partially skipped pipelines can still be merged.

License

This project is distributed under the "QQRM LAPOCHKA v1.0 License (AI-first Vibecoder)" in LICENSE_QQRM_LAPOCHKA. Contributors must generate changes via an AI agent. Manual code submissions may still be humorously called a "skill issue" by the community.