Scarp
हिन्दी

Download Scarp Docs for offline reading

The official Scarp Docs can also be packaged as an English/Hindi offline snapshot. A downloadable archive contains the built VitePress pages, styles, JavaScript, and local-search assets from one tested Git commit. It is documentation only, not a Scarp compiler download.

Download the tested archive

  1. Open Scarp Docs deployment workflow.
  2. Choose a successful run on main, preferably the latest. In its Artifacts section, select the scarp-docs-offline-… artifact. GitHub may require you to sign in.
  3. Extract the downloaded GitHub artifact ZIP. It contains scarp-docs-offline.tar.gz and scarp-docs-offline.tar.gz.sha256.
  4. Check the SHA-256 before extracting the tar archive. The hash file guards against accidental corruption, not against a compromised distribution channel.
  5. Extract the tar archive into a fresh folder. It includes README-OFFLINE.txt, MANIFEST.json, serve.mjs, and a scarp-docs/ directory.

These Actions artifacts have limited GitHub retention, unlike a permanent GitHub Release asset. If an old artifact has expired, choose a newer successful workflow run. No stable docs release or compiler binary is created by this workflow.

Verify the downloaded archive

On Linux, from the folder containing the archive and checksum:

sha256sum -c scarp-docs-offline.tar.gz.sha256
mkdir -p scarp-docs-extracted
tar -xzf scarp-docs-offline.tar.gz -C scarp-docs-extracted

On macOS, where sha256sum may not be installed, display the SHA-256 and compare it to the text in the .sha256 file:

shasum -a 256 scarp-docs-offline.tar.gz
mkdir -p scarp-docs-extracted
tar -xzf scarp-docs-offline.tar.gz -C scarp-docs-extracted

On Windows PowerShell, compare the file hash with the adjacent checksum file, then extract using Windows' built-in tar command:

Get-FileHash .\scarp-docs-offline.tar.gz -Algorithm SHA256
Get-Content .\scarp-docs-offline.tar.gz.sha256
New-Item -ItemType Directory -Path .\scarp-docs-extracted
tar -xzf .\scarp-docs-offline.tar.gz -C .\scarp-docs-extracted

Open the site locally

Install Node.js 20.19 or newer if you have not already. The archive does not need npm install or internet access to start its local server.

From the extracted scarp-docs-extracted folder, run:

node serve.mjs

Open the local addresses:

  • English: http://127.0.0.1:8080/scarp-docs/
  • हिन्दी: http://127.0.0.1:8080/scarp-docs/hi/

The included server binds only to 127.0.0.1 (this computer). Stop with Ctrl+C. Do not open index.html with a file:// URL: the generated site uses absolute /scarp-docs/ asset paths and requires HTTP serving.

What works offline?

Automated accessibility checks: The offline test suite scans a selection of actual English/Hindi pages and interactive states with locally bundled axe-core WCAG A/AA rules and checks 320px-wide reflow. These checks find common machine-detectable problems; they do not replace manual keyboard, screen-reader or contrast perception reviews.

Automated browser verification: the documentation CI now starts the extracted local server in headless Chromium and checks English/Hindi Command Finder interaction, real clipboard success and denied permissions, keyboard-activated local search and its empty-results state, language switching, skip-to-content, persisted dark/light appearance, mobile navigation, reduced-motion operation and the absence of external requests. The browser runs locally against the packaged archive, not the hosted GitHub Pages site. This does not constitute a complete WCAG audit.

The snapshot includes all generated English and Hindi pages, their local CSS/JavaScript, and the assets used by Scarp Docs local search and the version-aware Command Finder. Regular internal navigation runs against your local server. Links to GitHub, official releases, social profiles, or other external sites still require internet. The browser's actual keyboard/search behavior is being tracked separately in accessibility test issue #65; static asset checks are not equivalent to a full browser accessibility audit.

Read MANIFEST.json inside the extracted archive for the documentation version, exact docs source commit, compiler release boundary and SHA-256 digest of every packaged file. The core reference remains pinned to alpha.1; the latest separately documented published compiler is v0.0.6-alpha.1. An offline snapshot does not upgrade the interpreter.

For maintainers, the archive is generated by npm run package:offline after npm run docs:build; verify it using npm run test:offline. The checked Pages workflow uploads a fresh asset for each successful main build. हिन्दी offline guide covers the same process.

Search documentation