Reads as one continuous stream now. Each entry is just:
λ> <input>
<output> ; tier · NNms
No left border, no boxed cards, no side-column tier label. Output
indents under the prompt (3ch) using monospace ch units. Tier+time
render as a Lisp-comment-style suffix in muted color.
Prompt bar: borderless textarea on the code-bg surface so the input
visually joins the transcript above. Placeholder cut to "(+ 1 2)" —
the surrounding text already explains the semantics.
Multi-line inputs keep prompt continuation marks ("..").
124 lines
5.4 KiB
HTML
124 lines
5.4 KiB
HTML
<!DOCTYPE html>
|
|
<html lang="en">
|
|
<head>
|
|
<meta charset="utf-8">
|
|
<meta name="viewport" content="width=device-width, initial-scale=1">
|
|
<title>lumbda repl — three tiers, persistent sessions</title>
|
|
<link rel="stylesheet" href="style.css">
|
|
<link rel="stylesheet" href="repl.css">
|
|
<script type="importmap">
|
|
{
|
|
"imports": {
|
|
"@codemirror/state": "https://esm.sh/*@codemirror/state@6.4.1",
|
|
"@codemirror/view": "https://esm.sh/*@codemirror/view@6.34.1",
|
|
"@codemirror/commands": "https://esm.sh/*@codemirror/commands@6.6.2",
|
|
"@codemirror/language": "https://esm.sh/*@codemirror/language@6.10.3",
|
|
"@codemirror/legacy-modes/mode/scheme": "https://esm.sh/*@codemirror/legacy-modes@6.4.1/mode/scheme",
|
|
"@codemirror/theme-one-dark": "https://esm.sh/*@codemirror/theme-one-dark@6.1.2",
|
|
"@lezer/highlight": "https://esm.sh/*@lezer/highlight@1.2.1",
|
|
"@lezer/common": "https://esm.sh/*@lezer/common@1.2.3",
|
|
"style-mod": "https://esm.sh/*style-mod@4.1.2",
|
|
"crelt": "https://esm.sh/*crelt@1.0.6",
|
|
"w3c-keyname": "https://esm.sh/*w3c-keyname@2.2.8"
|
|
}
|
|
}
|
|
</script>
|
|
</head>
|
|
<body class="repl">
|
|
|
|
<header>
|
|
<a class="brand" href="../" aria-label="lumbda home">
|
|
<img class="logo" src="lumbda-logo-green.png" alt="" aria-hidden="true">
|
|
<h1 aria-label="lumbda.">lumbda<span class="period" aria-hidden="true">.</span></h1>
|
|
</a>
|
|
<p class="tagline">three tiers, persistent sessions</p>
|
|
<p class="lede">
|
|
interactive repl — each tab keeps its own tier state. <kbd>enter</kbd>
|
|
sends, <kbd>shift+enter</kbd> for newline. transcripts persist in
|
|
encrypted localStorage; supply a password and the same password
|
|
unlocks them after a reload. for one-shot demos, see the
|
|
<a href="../playground/">playground</a>.
|
|
</p>
|
|
</header>
|
|
|
|
<div id="lock-screen" class="lock-screen">
|
|
<div class="lock-card">
|
|
<h2>unlock</h2>
|
|
<p>type a password to unlock or create a vault. every keystroke and
|
|
transcript line stays inside your browser — nothing leaves
|
|
this tab.</p>
|
|
<details class="lock-explain">
|
|
<summary>how the encryption works</summary>
|
|
<ul>
|
|
<li>your password is run through <strong>PBKDF2</strong> (200,000
|
|
iterations, SHA-256) with a random 32-byte salt generated
|
|
once per device and stored in <code>localStorage</code>.</li>
|
|
<li>that derives an <strong>AES-GCM 256-bit key</strong>. all
|
|
transcripts are encrypted with this key before any
|
|
<code>localStorage.setItem()</code> call.</li>
|
|
<li>the <strong>vault id</strong> is
|
|
<code>SHA-256(password || salt)</code> — different
|
|
passwords on the same device address completely independent
|
|
vaults. wrong password = decryption fails, you don't see
|
|
other vaults' contents.</li>
|
|
<li>same protocol as <a href="https://unsandbox.com/blog/vault-encryption-design"
|
|
target="_blank">unsandbox's vault design</a>, implemented
|
|
on the native <code>crypto.subtle</code> API instead of a
|
|
shipped CryptoJS bundle. source:
|
|
<a href="crypto.js" target="_blank"><code>crypto.js</code></a>.</li>
|
|
<li>passwords are never sent anywhere — not to lumbda.com,
|
|
not to a backend, not to a worker. nothing to leak in
|
|
transit because nothing transits.</li>
|
|
<li><strong>ephemeral</strong> skips the vault entirely —
|
|
nothing is written to disk; closing the tab loses everything.
|
|
use this on a shared machine when persistence would be a
|
|
footgun.</li>
|
|
</ul>
|
|
</details>
|
|
<input id="password" type="password" autocomplete="off" autofocus placeholder="password">
|
|
<div class="lock-row">
|
|
<button id="unlock" class="primary">unlock vault</button>
|
|
<button id="fresh" class="secondary" title="skip the vault, run ephemerally — transcripts vanish on reload">ephemeral</button>
|
|
</div>
|
|
<p class="lock-note" id="lock-note"></p>
|
|
</div>
|
|
</div>
|
|
|
|
<section class="tabbar" id="tabbar" hidden>
|
|
<div class="tabs" id="tabs"></div>
|
|
<button id="new-tab" class="ghost" title="new repl session">+</button>
|
|
<span class="spacer"></span>
|
|
<button id="reset-tier" class="ghost" title="reboot the worker for the current tab — state lost">↻ reboot tier</button>
|
|
<button id="clear-log" class="ghost" title="clear the transcript">clear log</button>
|
|
<button id="cancel" class="ghost" disabled>cancel</button>
|
|
<button id="lock" class="ghost" title="lock vault — passes back to lock screen">lock</button>
|
|
</section>
|
|
|
|
<main id="transcript-wrap" class="transcript-wrap" hidden>
|
|
<div id="transcript" class="transcript"></div>
|
|
</main>
|
|
|
|
<section class="prompt-bar" id="prompt-bar" hidden>
|
|
<span class="prompt-sigil">λ></span>
|
|
<textarea id="input" rows="1" spellcheck="false"
|
|
placeholder="(+ 1 2)"></textarea>
|
|
<select id="tier-select" title="which tier evaluates this input">
|
|
<option value="python">python (pyodide)</option>
|
|
<option value="c" selected>c (emcc)</option>
|
|
<option value="asm">asm (wat)</option>
|
|
<option value="all">all three (race)</option>
|
|
</select>
|
|
<button id="send" class="primary">send</button>
|
|
</section>
|
|
|
|
<footer>
|
|
<p>
|
|
transcripts saved encrypted with your password. tier state (defines,
|
|
set!, hash table contents) lives in the worker; lost on reload. lift
|
|
a session via portal save when that lands in WAT. <a href="../">lumbda.</a>
|
|
</p>
|
|
</footer>
|
|
|
|
<script type="module" src="repl.js"></script>
|
|
</body>
|
|
</html>
|