<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Sudipto Chandra on Medium]]></title>
        <description><![CDATA[Stories by Sudipto Chandra on Medium]]></description>
        <link>https://medium.com/@dipu.sudipta?source=rss-4135668b097------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*22oODRcsNO_2tnV14BNQRg.jpeg</url>
            <title>Stories by Sudipto Chandra on Medium</title>
            <link>https://medium.com/@dipu.sudipta?source=rss-4135668b097------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 10:55:11 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@dipu.sudipta/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[What Happens When AI Grants Us the Universe?]]></title>
            <link>https://medium.com/@dipu.sudipta/what-happens-when-ai-grants-us-the-universe-a7e7bcb66b6b?source=rss-4135668b097------2</link>
            <guid isPermaLink="false">https://medium.com/p/a7e7bcb66b6b</guid>
            <category><![CDATA[agi]]></category>
            <category><![CDATA[thought-experiment]]></category>
            <category><![CDATA[future]]></category>
            <category><![CDATA[sia]]></category>
            <dc:creator><![CDATA[Sudipto Chandra]]></dc:creator>
            <pubDate>Thu, 29 Oct 2026 01:44:04 GMT</pubDate>
            <atom:updated>2026-10-29T01:44:04.863Z</atom:updated>
            <content:encoded><![CDATA[<h3><strong>How much would you pay for immortality, or even just to live 200 years of healthy life?</strong></h3><p>AI may make that possible within next few years. That possibility alone is enough to keep wealthy people investing in it. If not for themselves, for their families. The technology is hard to slow down at this point. What comes afterwards is the point that matters and worth thinking about right now.</p><p>Most major inventions in the history have made our life a little bit more comfortable, starting from fire to electricity. But this comfort comes at a price. Some people gained immense benefits and some are left worse off. The gains never arrive evenly. Over time, though, technology has widened what people are able to do. AI will probably work the same way.</p><p>Of all the possible futures I can imagine, I like the one where people and machines grow closer together (Cyberpunk 2077, more or less). Humans augment themselves and Machines copy and improve themselves. Energy and labour stop being scarce. People get more time to enjoy life. Running of the world happens in the background.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*giU4QqL1bPWQFGnE" /></figure><h3><strong>What happens after machines take over?</strong></h3><p>For most of our history, survival took nearly all our attention. Just couple of centuries ago, people lived in the forest not knowing whether they can eat and survive the tomorrow. We have built technology to free ourselves from the rough side of mother nature.</p><p>AGI is going to be just a new kind of technology that will free all of us from our mandatory labours. is a If you have more freedom, does it automatically mean peace? Of course not. If people have more time and more security, the population may grow. Growth puts pressure on land and resources, and pressure turns into conflict. The 22nd century can become a period of war. It may not be an all-out war, but a silent one, where people can die without uttering a single word.</p><p>This war could last until we find somewhere else to live, another planet to pull resources from. It is just a matter of how long the machines will take to find a way to colonise a planet. In the process of gaining back the stability, the meaning of being a human may change significantly.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*EzIG1O7c7UTstDLB" /></figure><h3><strong>What comes afterwards?</strong></h3><p>We will proceed to colonise the entire solar system over the next 500 years. Then run into the same fight over resources on a bigger stage. The way out of this is to go to another star. We will need a self-sustainable ship large enough for interstellar travel. Machines will eventually build that and the entire galaxy will open up to us. We will get access to a field so big that filling it up will take millions of years.</p><p>This is one of the most optimistic future. It leaves out everything we have not invented yet, and challenges we have yet to comprehend. We will have to just trust our machine god blindly, and pray that laws of physics will not put a wall in front of it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*n1GKf_Mm-xvtuDKK" /></figure><h3><strong>Will we survive?</strong></h3><p>It is not the laws of physics that worries me, but the human nature itself. There is no cure for the <a href="https://en.wikipedia.org/wiki/Dark_triad">Dark Triad</a>. No advanced, rational entity should have an obvious reason to let primitive humans keep breeding forever. We will either go extinct or evolve into something else.</p><p>After generations of evolution, we may loose what makes us human eventually. Our identity as human will become another <a href="https://en.wikipedia.org/wiki/Ship_of_Theseus">Theseus’s paradox</a>. Whether you call that the survival of humanity is a matter of taste.</p><p>On the bright side, whatever we will turn into may not need millions of years to cross a galaxy, a several thousands might be enough before we start another war to conquer the entire Universe.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a7e7bcb66b6b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Bypassing Cloudflare and DDoS-Guard: A Complete Technical Reference for Web Scrapers (2026)]]></title>
            <link>https://medium.com/@dipu.sudipta/bypassing-cloudflare-and-ddos-guard-a-complete-technical-reference-for-web-scrapers-2026-6375c38af658?source=rss-4135668b097------2</link>
            <guid isPermaLink="false">https://medium.com/p/6375c38af658</guid>
            <category><![CDATA[automation]]></category>
            <category><![CDATA[web-scraping]]></category>
            <category><![CDATA[cloudflare]]></category>
            <category><![CDATA[web-application-firewall]]></category>
            <dc:creator><![CDATA[Sudipto Chandra]]></dc:creator>
            <pubDate>Tue, 09 Jun 2026 13:35:29 GMT</pubDate>
            <atom:updated>2026-06-10T07:31:04.184Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*P0mqg1STNqdeCMX8.jpg" /></figure><p><strong><em>Scope:</em></strong><em> This article covers every known detection layer used by Cloudflare and DDoS-Guard, the tools and techniques available to bypass each one, and the layers that currently have no reliable bypass. It is written for developers building legitimate web scrapers and data pipelines who need to understand the full landscape of modern bot protection in 2026.</em></p><p>The bot detection industry shifted from simple IP blocklists to <strong>composite trust scoring</strong> between 2022 and 2026. Modern systems like Cloudflare no longer make a binary “bot or not” decision on a single signal. Instead, they compute a score from 6–8 independent detection layers simultaneously. A scraper only gets through if it passes <em>all active layers at once.</em> Failing even one is usually enough to trigger a challenge or block.</p><p>This means there is no silver bullet. A tool that fixes TLS fingerprinting but ignores behavioral signals will still get blocked. Understanding each layer independently is the prerequisite to building a working bypass stack.</p><figure><img alt="Bypassing Cloudflare" src="https://cdn-images-1.medium.com/max/1024/0*UzOBtrVGEaY8K3mT" /></figure><blockquote>Cloudflare protects an estimated 74–79% of all websites that use any bot mitigation tooling, making it by far the most commonly encountered wall in web scraping. Its detection stack has grown substantially more sophisticated since 2024.</blockquote><h3>Layer 1 — IP Reputation</h3><h4><strong>How it works</strong></h4><p>Every incoming IP is scored against Cloudflare’s global threat intelligence database. The score incorporates: known datacenter ASN ranges (AWS, GCP, Azure, Hetzner, DigitalOcean, etc.), Tor exit node lists, residential proxy pool reputation (shared flagged history across users), and CGNAT classification for mobile carrier IPs.</p><h4><strong>Bypass options</strong></h4><p>The effectiveness of different IP types (as of 2026) are presented in the following table:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fdatawrapper.dwcdn.net%2FB4Nl1%2F3%2F&amp;display_name=Datawrapper&amp;url=https%3A%2F%2Fdatawrapper.dwcdn.net%2FB4Nl1%2F3%2F&amp;image=https%3A%2F%2Fdatawrapper.dwcdn.net%2FB4Nl1%2Fplain-s.png%3Fv%3D3&amp;type=text%2Fhtml&amp;schema=dwcdn" width="700" height="378" frameborder="0" scrolling="no"><a href="https://medium.com/media/ff44769a48dcef5ebb69183acdff771f/href">https://medium.com/media/ff44769a48dcef5ebb69183acdff771f/href</a></iframe><p><strong>Recommended approach:</strong> Mobile carrier proxies for high-security targets; residential proxies for medium-difficulty targets; never use datacenter IPs for Cloudflare-protected scraping.</p><h3>Layer 2 — TLS / JA3 / JA4 Fingerprinting</h3><h4><strong>How it works</strong></h4><p>When any HTTP client connects over HTTPS, it sends a TLS ClientHello handshake. Cloudflare extracts a fingerprint from this handshake and compares it against a database of known browser profiles.</p><ul><li><strong>JA3</strong> (2017) — MD5 hash of TLS version, cipher suites, extensions, elliptic curves, and elliptic curve formats.</li><li><strong>JA4</strong> (2023, now standard) — Multi-part fingerprint covering TLS version, SNI presence, cipher count, extension count, ALPN, and separate hashes for ciphers and extensions. More resistant to evasion than JA3. Cloudflare analyzes over 15 million unique JA4 fingerprints daily.</li></ul><p>Python’s requests library, httpx, and most HTTP clients produce a TLS handshake that looks nothing like Chrome. Cloudflare identifies the mismatch in milliseconds — not because of what you requested, but because of <em>how</em> you connected.</p><h4><strong>The 2026 addition — Post-Quantum TLS (ML-KEM)</strong></h4><p>Chrome 116+ sends a post-quantum key share (X25519Kyber768 / ML-KEM) as part of its ClientHello. Cloudflare now tracks this as part of the fingerprint. A scraper that claims to be Chrome but omits the PQ key share produces a fingerprint that matches no known-good browser profile, creating an anomaly signal.</p><h4><strong>Bypass options</strong></h4><ul><li><strong>curl_cffi (Python)</strong> — The most practical Python solution. Wraps curl-impersonate to reproduce exact browser TLS signatures including cipher suite order, extensions, ALPN, and post-quantum key shares when using recent Chrome profiles. Drop-in replacement for requests:</li></ul><pre>from curl_cffi import requests<br>response = requests.get(url, impersonate=&quot;chrome124&quot;)</pre><p>Keep the impersonate target pinned to a recent Chrome version (124+) to include the PQ key share.</p><ul><li><strong>tls-client (Python/Go)</strong> — More fingerprint profiles than curl_cffi, slightly less actively maintained.</li><li><strong>cycletls (Go)</strong> — Excellent for Akamai-grade JA4 spoofing; preferred in Go-based scraping pipelines.</li><li><strong>got-scraping (Node.js)</strong> — Native Node.js solution with automatic header generation.</li></ul><h4><strong>What does NOT work</strong></h4><ul><li>Raw requests or httpx — Both produce non-browser TLS signatures. httpx with http2=True fixes the protocol version but not the cipher suite order or extension fingerprint.</li><li>Any HTTP client that doesn’t implement curl-impersonate under the hood.</li></ul><h3>Layer 3 — HTTP/2 Fingerprinting</h3><h4><strong>How it works</strong></h4><p>Real browsers send HTTP/2 with a specific frame ordering: SETTINGS → WINDOW_UPDATE → HEADERS. The SETTINGS frame contains specific parameter values (header table size, initial window size, max concurrent streams) that match the browser version. Cloudflare’s heuristics engine includes approximately 50 hand-written rules targeting HTTP/2 frame fingerprints.</p><p>Python requests defaults to HTTP/1.1. Even httpx with HTTP/2 enabled sends non-browser frame ordering.</p><h4><strong>Bypass options</strong></h4><ul><li><strong>curl_cffi</strong> — Fixes both TLS and HTTP/2 frame ordering simultaneously with the same impersonate parameter. This is the main reason it is the recommended tool.</li><li><strong>Real browser</strong> (any) — Sends correct HTTP/2 by definition.</li></ul><p><strong>Common mistake:</strong> Enabling HTTP/2 in httpx or aiohttp does not fix this. The frames are sent, but in the wrong order with wrong parameter values. The TLS + HTTP/2 combination must match a specific browser version exactly.</p><h3>Layer 4 — Browser / JavaScript Fingerprinting</h3><h4><strong>How it works</strong></h4><p>When a challenge page or protected page loads, Cloudflare’s injected JavaScript collects device fingerprints including:</p><ul><li><strong>Canvas fingerprint</strong> — Renders a hidden canvas element and hashes the output. GPUs produce slightly different outputs per hardware/driver combination. Headless browsers produce consistent, suspiciously uniform outputs.</li><li><strong>WebGL fingerprint</strong> — GPU renderer string and vendor. Headless Chrome reports “SwiftShader” or “Mesa” — immediate bot signals.</li><li><strong>Audio context fingerprint</strong> — Processes a short audio buffer through the AudioContext API. Results vary by hardware.</li><li><strong>Font enumeration</strong> — Checks for system fonts that correlate with the claimed OS.</li><li><strong>navigator.webdriver</strong> — Set to true in all standard Playwright/Selenium/Puppeteer instances. Cloudflare checks this directly.</li><li><strong>Permissions API anomalies</strong> — Headless browsers handle notification/camera/microphone permission queries differently than real browsers.</li><li><strong>Plugin list</strong> — Real Chrome has plugins; headless Chrome has none.</li><li><strong>Screen dimensions and color depth</strong> — Headless defaults are often unset or inconsistent.</li></ul><h4><strong>Bypass options</strong></h4><ul><li><strong>Camoufox</strong> — Firefox-based stealth browser. Actively maintained as of 2026. Patches dozens of fingerprinting signals. Performs well on targets where Chrome-shaped fingerprints are specifically being profiled. Uses Playwright’s API internally so the integration surface is familiar:</li></ul><pre>from camoufox.sync_api import Camoufox<br>with Camoufox(headless=False) as browser:<br>    page = browser.new_page()<br>    page.goto(url)<br>    html = page.content()</pre><ul><li><strong>Patchright</strong> — Chromium-based fork of Playwright with 49+ C++ patches targeting fingerprint anomalies. Better when Chrome fingerprints are required specifically.</li><li><strong>SeleniumBase UC Mode</strong> — Patches navigator.webdriver and several other common detection points. Good for mid-difficulty targets; free and lightweight.</li></ul><h4><strong>What does NOT work</strong></h4><ul><li>playwright-extra with the stealth plugin — Deprecated by its maintainers in February 2025. Current Cloudflare fingerprinting detects it reliably.</li><li>Standard Playwright or Puppeteer without patching — Flagged immediately on navigator.webdriver alone.</li></ul><h3>Layer 5 — CDP / DevTools Protocol Detection</h3><h4><strong>How it works</strong></h4><p>Chrome DevTools Protocol (CDP) is the interface that Playwright, Puppeteer, Selenium, and all similar tools use to control a browser. Cloudflare’s JavaScript probes for CDP artifacts that persist even after common patches like setting navigator.webdriver = false. These include:</p><ul><li>Specific property descriptors on window and document that CDP modifies</li><li>Timing anomalies in how certain JS APIs respond when driven by automation</li><li>Inconsistencies in how Runtime.evaluate and Page.addScriptToEvaluateOnLoad modify the JS environment</li></ul><h4><strong>Bypass options</strong></h4><ul><li><strong>nodriver</strong> — Successor to undetected-chromedriver. Implements its own custom DevTools interface instead of using the standard CDP automation protocol that detectors target. Currently the best option for automation-protocol-level detection:</li></ul><pre>import nodriver as uc<br><br>async def main():<br>    browser = await uc.start(headless=False)<br>    page = await browser.get(url)<br>    html = await page.get_content()<br>    browser.stop()<br><br>uc.loop().run_until_complete(main())</pre><ul><li><strong>Headful mode (</strong><strong>headless=False)</strong> — Always produces better results than headless. GPU renderer, screen properties, and several other signals only look correct in headful mode.</li></ul><p><strong>Partially unresolvable:</strong> Camoufox and Patchright both use Playwright’s automation protocol internally, so they remain detectable by automation-protocol fingerprinting even with all their other patches applied. nodriver is the only freely available tool that meaningfully addresses this specific layer.</p><h3>Layer 6 — Behavioral / ML Signals</h3><h4><strong>How it works</strong></h4><p>Since 2025, Cloudflare deploys per-zone ML anomaly models — meaning the model is trained separately per website, not shared globally. What looks like normal behavior on a developer documentation site may be flagged as anomalous on an e-commerce checkout flow. Signals include:</p><ul><li><strong>Request timing regularity</strong> — Perfectly timed requests are anomalous. Human browsing has irregular timing.</li><li><strong>Navigation patterns</strong> — Real users follow referrer chains (homepage → listing → detail). Scrapers often jump directly to deep URLs.</li><li><strong>Scroll and mouse events</strong> — On JS-rendered pages, bots that don’t scroll or move the mouse before reading content are distinguishable.</li><li><strong>Cookie and session age</strong> — Fresh sessions with no history are higher risk than returning sessions with accumulated cookies.</li><li><strong>Inter-request intervals</strong> — Sub-second intervals between page loads are a strong bot signal.</li></ul><h4><strong>Bypass options</strong></h4><ul><li>Add time.sleep(random.uniform(1.5, 5.0)) between requests. Vary the distribution — exponential or gamma distributions feel more natural than uniform.</li><li>Session warm-up: visit the homepage first, wait for JS to execute, accumulate cookies, then navigate to target pages.</li><li>Pass realistic referrer headers matching natural navigation (e.g. Referer: https://site.com/listing/ when scraping a detail page).</li><li>Persist cookies and session state across requests within a scraping session. Never start fresh on every connection.</li><li>For browser-based scrapers, add randomized mouse movement before any interaction.</li></ul><p><strong>Limitation:</strong> Per-zone ML makes this a moving target. A behavioral pattern that works on one Cloudflare-protected site may not work on another. Calibrating per target is necessary for high-value scraping.</p><h3>Layer 7 — Managed Challenge (JavaScript Challenge)</h3><h4><strong>How it works</strong></h4><p>When Cloudflare’s composite score is uncertain — not confident enough to block, not confident enough to allow — it serves an interstitial “Checking your browser…” page. This page executes JavaScript, verifies the client can run it, checks the resulting fingerprint, and if satisfied issues a cf_clearance cookie that grants session access.</p><h4><strong>Bypass options</strong></h4><p><strong><em>Option A — Cookie reuse (most efficient)<br></em></strong>Solve the challenge once with a full stealth browser, extract the resulting cookies, and pass them to a lightweight HTTP client for all subsequent requests. The cf_clearance cookie is typically valid for 30 minutes to 2 hours:</p><ol><li>Launch nodriver or Camoufox, navigate to the target site</li><li>The browser executes the challenge natively and receives cf_clearance</li><li>Extract cf_clearance and __cf_bm cookies plus the browser&#39;s User-Agent</li><li>Pass these cookies to curl_cffi with matching impersonate profile and User-Agent</li><li>Make all subsequent requests with curl_cffi — no browser overhead on every request</li></ol><p><strong>Critical:</strong> cf_clearance is bound to the IP address it was issued on. Use the same proxy when solving and when reusing.</p><p><strong><em>Option B — Solve inline<br></em></strong>Use nodriver or Camoufox for every request. Higher resource cost but simpler pipeline. Appropriate when session counts are low or when JS execution is required for the page content anyway.</p><h4><strong>What does NOT work</strong></h4><ul><li><strong>FlareSolverr</strong> — Deprecated and now easily detected by Cloudflare. Do not use.</li><li><strong>cloudscraper</strong> — Effectively dead. Cloudflare&#39;s challenge format has moved on.</li></ul><h3>Layer 8 — Cloudflare Turnstile CAPTCHA</h3><h4><strong>How it works</strong></h4><p><strong>Turnstile</strong> replaced <strong>hCaptcha</strong> as Cloudflare’s user-facing CAPTCHA. It runs a client-side JS challenge that evaluates browser authenticity, canvas fingerprint, and interaction history. It sometimes presents a visible checkbox; on suspicious sessions, it presents an image challenge. Unlike a simple JS check, Turnstile actively evaluates the quality of the browser environment before issuing a token.</p><h4><strong>Bypass options</strong></h4><p><strong><em>Option A — Click-based solver (free, less reliable)<br></em></strong>A stealth browser (Camoufox or Patchright) automatically clicks the checkbox. No external service needed. Works best on lower-sensitivity Turnstile configurations. Standard Playwright without patching will be detected and blocked before it reaches the click.</p><p><strong><em>Option B — CAPTCHA solving service (paid, more reliable)</em><br></strong>Services like 2Captcha and CapSolver accept the Turnstile challenge parameters, solve it via real humans or AI, and return a valid token. Cost is approximately $2–4 per 1,000 solves with a 10–30 second latency per solve. CapSolver’s AI system adapts to Cloudflare updates faster than most DIY implementations.</p><p><strong><em>Option C — Token reuse</em><br></strong>Turnstile tokens are valid for approximately 5 minutes. For scraping multiple pages in a session, solve once and reuse the token across subsequent requests.</p><h3>Layer 9 — Bot Fight Mode (Free Tier)</h3><h4><strong>How it works</strong></h4><p>Cloudflare’s free-tier bot protection. Uses simple heuristics rather than ML: known bad IP lists, missing or inconsistent HTTP headers (Accept-Language, Accept-Encoding), obviously automated User-Agents, datacenter IP ranges, and absence of browser cookies. It does not run full TLS fingerprinting or behavioral analysis.</p><h4><strong>Bypass options</strong></h4><p>This is the easiest Cloudflare layer to bypass. The combination of the following tools will pass Bot Fight Mode with very high reliability:</p><ul><li>curl_cffi with impersonate=&quot;chrome124&quot; (fixes TLS + headers)</li><li>Any residential or mobile IP</li><li>Standard browser headers</li></ul><p>Most public websites on Cloudflare’s free plan are only protected by this layer, making curl_cffi + residential proxy the right default starting point.</p><h3>Layer 10 — ML Bot Score v9 (Paid Bot Management)</h3><h4><strong>How it works</strong></h4><p>Available to Cloudflare paying customers. The v9 model (default on new zones from late 2025) is a composite ML score trained on all previous layers simultaneously — TLS fingerprint, HTTP/2 fingerprint, browser JS signals, behavioral patterns, IP reputation, session history, and cross-zone threat intelligence. A site using this is far more difficult to scrape reliably than one on the free tier.</p><h4><strong>Bypass options</strong></h4><p>No single tool bypasses this. The minimum working stack:</p><ol><li><strong>Mobile carrier IP</strong> — Non-flagged, CGNAT-protected</li><li><strong>curl_cffi with Chrome 124+</strong> — TLS, HTTP/2, and PQ key share all correct</li><li><strong>cf_clearance from nodriver or Camoufox</strong> — Valid session cookie from a passing browser challenge</li><li><strong>Session warming</strong> — Cookies accumulated, referrer chain intact, realistic delays</li><li><strong>All request headers matching the impersonated browser version exactly</strong></li></ol><p>With this full stack, success rates are high for light and medium Bot Management configurations. For sites with aggressive settings or behavioral ML tuning, success is inconsistent. These targets are typically high-value commercial sites (large e-commerce, financial platforms) rather than general content sites.</p><h3>Layer 11 — AI Bot Blocker (2024+)</h3><h4><strong>How it works</strong></h4><p>In July 2024, Cloudflare launched a one-click toggle (available on all plans including free) to block known AI crawler bots. It targets specific User-Agents associated with AI scrapers: GPTBot, ClaudeBot, CommonCrawl, CCBot, and others.</p><h4><strong>Bypass</strong></h4><p>Use a standard browser User-Agent. This feature targets specifically labeled AI crawlers, not general scrapers with browser User-Agents. Trivially avoided.</p><h3>Layer 12 — Web Bot Auth (May 2026 — NEW)</h3><h4><strong>How it works</strong></h4><p>The most significant new Cloudflare feature in 2026. Web Bot Auth is a cryptographic agent passport system. Sites that opt in can require connecting bots to present a cryptographically signed credential proving they are a registered, authorized agent. Partnered initially with Anchor Browser and Browserbase.</p><h4><strong>Bypass: ❌ Not possible</strong></h4><p>This is cryptographic authentication. There is no way to forge a credential you do not legitimately hold. If a site enables Web Bot Auth, unauthorized scrapers are blocked at the protocol level regardless of how convincing the rest of the fingerprint stack is.</p><p>The only path through is legitimate registration as an authorized agent. Web Bot Auth is brand new (May 2026), adoption is currently low, and it is opt-in — so it is not yet a common wall. However, it represents the likely future trajectory of bot protection for high-value targets.</p><h3>Layer 13 — Cloudflare Access (Zero Trust)</h3><h4><strong>How it works</strong></h4><p>Login-gated content behind Cloudflare’s identity provider. Requires authentication via SSO, OAuth, or email OTP before any page content is accessible.</p><h4><strong>Bypass: ❌ Not applicable</strong></h4><p>This is authentication, not anti-bot protection. Bypassing it without credentials would constitute unauthorized access. Not a web scraping scenario — if content requires login, the appropriate path is authenticated scraping with a legitimate account.</p><h3>Cloudflare Bypass Summary Table</h3><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fdatawrapper.dwcdn.net%2FdGZmH%2F1%2F&amp;display_name=Datawrapper&amp;url=https%3A%2F%2Fdatawrapper.dwcdn.net%2FdGZmH%2F1%2F&amp;image=https%3A%2F%2Fdatawrapper.dwcdn.net%2FdGZmH%2Fplain-s.png%3Fv%3D1&amp;type=text%2Fhtml&amp;schema=dwcdn" width="700" height="688" frameborder="0" scrolling="no"><a href="https://medium.com/media/b50a56664591a03d484cccc01017436b/href">https://medium.com/media/b50a56664591a03d484cccc01017436b/href</a></iframe><figure><img alt="" src="https://cdn-images-1.medium.com/max/716/1*u3ulLyxUe3qY7p6ztSFwHA.jpeg" /></figure><blockquote>DDoS-Guard is significantly less sophisticated than Cloudflare but holds approximately 3.5% of the bot mitigation market (globally) with dominant presence in Russian and Eastern European web infrastructure. It is the second most commonly encountered protection in the wild after Cloudflare.</blockquote><h3>Detection Layers</h3><ul><li><strong>IP Reputation</strong> Basic datacenter and Tor blocking. Much less aggressive than Cloudflare’s ASN-level profiling. Residential proxies work reliably.</li><li><strong>JavaScript Challenge</strong> Functionally similar to Cloudflare’s managed challenge — serves an interstitial page that runs JS and issues a session cookie. Notably, <strong>FlareSolverr still works against DDoS-Guard</strong> (unlike Cloudflare, where it has been defunct since early 2025). Tools like Camoufox and nodriver also solve it natively.</li><li><strong>TLS Fingerprinting</strong> Basic JA3/JA4 checking. Less comprehensive than Cloudflare’s implementation. curl_cffi with any impersonate profile resolves this without issue.</li><li><strong>WAF Rules</strong> Signature-based blocking of known bad request patterns. Resolved by using proper browser headers and a residential IP.</li><li><strong>Rate Limiting</strong> Per-IP request frequency limits. Standard rotating proxy + throttled request rate is sufficient.</li><li><strong>Behavioral ML</strong> Significantly weaker than Cloudflare. Random delays between requests are generally sufficient.</li></ul><h3>DDoS-Guard Bypass Stack</h3><p>For the vast majority of DDoS-Guard-protected targets, the following is sufficient:</p><ol><li>curl_cffi with impersonate=&quot;chrome124&quot; — handles TLS + headers</li><li>Residential proxy — passes IP reputation</li><li>FlareSolverr OR session cookie from Camoufox — handles the JS challenge</li><li>Random delays of 1–3 seconds between requests — handles rate limiting</li></ol><p>DDoS-Guard does not require the full multi-layer stack that Cloudflare’s advanced Bot Management demands.</p><h4>DDoS-Guard Bypass Summary Table</h4><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fdatawrapper.dwcdn.net%2Fsv5Ha%2F2%2F&amp;display_name=Datawrapper&amp;url=https%3A%2F%2Fdatawrapper.dwcdn.net%2Fsv5Ha%2F2%2F&amp;image=https%3A%2F%2Fdatawrapper.dwcdn.net%2Fsv5Ha%2Fplain-s.png%3Fv%3D2&amp;type=text%2Fhtml&amp;schema=dwcdn" width="600" height="300" frameborder="0" scrolling="no"><a href="https://medium.com/media/a33cca240c596375eb043b634e50dbaa/href">https://medium.com/media/a33cca240c596375eb043b634e50dbaa/href</a></iframe><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*aL0Q4j0xDDuMaY26Fom7vw.png" /></figure><h3>Recommended Tool Stack by Scenario</h3><h4>Scenario A — Basic protection (Bot Fight Mode / DDoS-Guard / no JS challenge)</h4><pre>from curl_cffi import requests<br><br>session = requests.Session()<br>response = session.get(<br>    url,<br>    impersonate=&quot;chrome124&quot;,<br>    headers={<br>        &quot;User-Agent&quot;: &quot;Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36&quot;,<br>        &quot;Accept-Language&quot;: &quot;en-US,en;q=0.9&quot;,<br>        &quot;Accept&quot;: &quot;text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8&quot;,<br>    }<br>)</pre><p>Covers roughly 70–80% of protected sites encountered in practice.</p><h4>Scenario B — Managed challenge / Turnstile (standard Cloudflare)</h4><p>Solve once, reuse cookie pattern:</p><pre>import nodriver as uc<br>from curl_cffi import requests as cffi_requests<br><br># Step 1: Solve with real browser<br>async def get_clearance(url, proxy):<br>    browser = await uc.start(<br>        headless=False,<br>        browser_args=[f&quot;--proxy-server={proxy}&quot;]<br>    )<br>    page = await browser.get(url)<br>    await page.sleep(3)  # wait for challenge resolution<br>    cookies = await page.browser.cookies([url])<br>    cf_clearance = next(c[&quot;value&quot;] for c in cookies if c[&quot;name&quot;] == &quot;cf_clearance&quot;)<br>    browser.stop()<br>    return cf_clearance<br><br># Step 2: Reuse cookie in lightweight client<br>def scrape_with_clearance(url, cf_clearance, proxy):<br>    session = cffi_requests.Session()<br>    session.cookies.set(&quot;cf_clearance&quot;, cf_clearance)<br>    return session.get(<br>        url,<br>        impersonate=&quot;chrome124&quot;,<br>        proxies={&quot;https&quot;: proxy},<br>        headers={&quot;User-Agent&quot;: &quot;Mozilla/5.0 ...&quot;}<br>    )</pre><h4>Scenario C — Full Bot Management / high-security target</h4><p>Use a managed scraping API (ScrapFly, BrightData, ScrapingBee). At this tier, maintaining a DIY bypass stack is effectively a full-time job given how frequently Cloudflare updates its detection. The per-request cost of a managed API ($0.003–$0.01 per request) is almost always cheaper than the engineering time required to maintain reliable DIY bypasses on ML Bot Score v9 targets.</p><h3>What No Longer Works in 2026</h3><p>A summary of tools and techniques that were effective in prior years but are now reliably detected:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fdatawrapper.dwcdn.net%2FdTQZH%2F1%2F&amp;display_name=Datawrapper&amp;url=https%3A%2F%2Fdatawrapper.dwcdn.net%2FdTQZH%2F1%2F&amp;image=https%3A%2F%2Fdatawrapper.dwcdn.net%2FdTQZH%2Fplain-s.png%3Fv%3D1&amp;type=text%2Fhtml&amp;schema=dwcdn" width="700" height="538" frameborder="0" scrolling="no"><a href="https://medium.com/media/25496e9f26b8f4ec5ce31131b4efebb4/href">https://medium.com/media/25496e9f26b8f4ec5ce31131b4efebb4/href</a></iframe><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*05xrKTD6xmIgmPi2.png" /></figure><h3>Legal and Ethical Considerations</h3><p>Bypassing bot detection is a technical capability, not a legal right. Before using any of these techniques, consider:</p><ul><li><strong>Terms of Service</strong> — Most websites prohibit automated access. Violating ToS may expose you to civil liability depending on jurisdiction.</li><li><strong>Computer fraud laws</strong> — In some jurisdictions (CFAA in the US, Computer Misuse Act in the UK), accessing a system in a manner that circumvents technical protection measures may carry legal risk even for publicly accessible content.</li><li><strong>Rate and impact</strong> — Even technically permitted scraping can become tortious if the request volume materially impairs the target server.</li><li><strong>Data protection</strong> — If scraped data includes personal information (names, emails, profiles), GDPR, CCPA, and similar regulations may apply to how that data is stored and used.</li></ul><p>The techniques in this article are documented for developers engaged in legitimate data collection — price monitoring, research, content aggregation of public information, accessibility tools, and similar use cases. Where an official API exists, use it.</p><h3>Conclusion</h3><p>Modern bot protection in 2026 is a multi-layer composite system, not a single wall. Cloudflare’s strength lies in requiring bypass of <em>all</em> layers simultaneously — TLS fingerprint, HTTP/2 frames, browser JS environment, CDP protocol, behavioral signals, and IP reputation all need to be consistent and convincing at the same time.</p><p>The practical takeaway:</p><ul><li><strong>curl_cffi</strong> with a current Chrome impersonation profile is the baseline starting point for any anti-bot bypass work. It handles TLS, HTTP/2, and headers in one call.</li><li><strong>nodriver</strong> is the strongest available tool for JS challenges and CDP detection.</li><li><strong>Camoufox</strong> is the best option when Firefox fingerprints are needed or Chrome-based tools are specifically being targeted.</li><li><strong>Mobile carrier IPs</strong> provide the best IP reputation foundation in 2026.</li><li><strong>Cookie reuse</strong> (solve once with browser, execute many with curl_cffi) is the most resource-efficient pattern for handling managed challenges at scale.</li><li><strong>Web Bot Auth</strong> and <strong>Cloudflare Access</strong> are the two layers with no current bypass — the former by cryptographic design, the latter by authentication requirement.</li></ul><p>DDoS-Guard, while less sophisticated, is still meaningfully protective and is the second most commonly encountered protection. It is effectively resolved by curl_cffi + residential proxy + FlareSolverr for the JS challenge layer.</p><p><em>Last updated: June 2026. Bot detection systems update frequently; specific tool versions and success rates should be verified against current community benchmarks before production deployment.</em></p><p><em>A more up to date and complete version of this artcle can be found here: </em><a href="https://github.com/lncrawl/scraper/wiki/Bypassing-Cloudflare"><em>https://github.com/lncrawl/scraper/wiki/Bypassing-Cloudflare</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6375c38af658" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Remotely access an intranet server through a proxy via an SSH tunnel from a shared terminal over…]]></title>
            <link>https://medium.com/@dipu.sudipta/remotely-access-an-intranet-server-through-a-proxy-via-an-ssh-tunnel-from-a-shared-terminal-over-aee9b4c2c25e?source=rss-4135668b097------2</link>
            <guid isPermaLink="false">https://medium.com/p/aee9b4c2c25e</guid>
            <category><![CDATA[terminal]]></category>
            <category><![CDATA[devops]]></category>
            <category><![CDATA[remote-access]]></category>
            <category><![CDATA[docker]]></category>
            <category><![CDATA[ssh]]></category>
            <dc:creator><![CDATA[Sudipto Chandra]]></dc:creator>
            <pubDate>Fri, 12 Jul 2019 16:19:35 GMT</pubDate>
            <atom:updated>2026-06-10T07:35:45.528Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/650/1*8dSS8ozKADNR2Iz_iYy73w.jpeg" /></figure><h3>Remotely access an intranet server through a proxy via an SSH tunnel with a web-terminal</h3><p>Most software companies use their own internal network or intranet, where all there computers are connected. Usually, a few among those machines are given access from the internet. But what if it is required to access to a computer on this network that is not reachable from the internet?</p><figure><img alt="Figure of some servers on the local network connected to the internet via a proxy server" src="https://cdn-images-1.medium.com/max/601/1*dTbxQLcxtCgSEyZLSPQX5w.jpeg" /><figcaption><strong>Figure-1:</strong> Servers on the local network are connected to the internet via a proxy server.</figcaption></figure><p>Consider the network topology presented in <strong>Figure-1</strong>. The <em>proxy server </em>here is accessible over the internet. Other servers on the network are connected to the <em>proxy server,</em> but not directly reachable from the internet. If some <em>client computers </em>try to connect to the any of the <em>target server, </em>the connection has to go via the <em>proxy server.</em> Lets find if there is any way to do that using the current technology available on the internet.</p><h4>Tools that we will need</h4><p>To do this experiment, we need these tools:</p><ul><li><a href="https://github.com/tsl0922/ttyd">TTYD</a> (a simple command-line tool for sharing terminal over the web)</li><li>OpenSSH Server</li><li>Docker</li></ul><h4>How to do it</h4><p>First you have to enable the ssh on all of the <em>target </em>and <em>proxy </em>servers, if it is already not available.</p><blockquote><em>On Ubuntu:<br></em><em>apt install openssh-server</em></blockquote><p>Now generate SSH key-pairs using:</p><blockquote>ssh-keygen -t rsa -C &quot;&lt;your email address&gt;&quot; <br>You will be prompt to enter the location of .ssh folder and a password for the private key. Just press Enter to use default configs in both times.</blockquote><p>Next, setup Docker on the <em>proxy </em>server<em>.</em></p><blockquote><em>On Ubuntu:</em><br><em>apt install docker-ce</em></blockquote><p>Run the <strong>ttyd</strong> server using docker on <em>proxy </em>server.</p><blockquote><em>docker run -itd --rm -p 7681:7681 tsl0922/ttyd</em></blockquote><p>From your client computer, now you can access the <em>ttyd</em> terminal. Open the browser and enter the proxy server’s host address and 7681 as the port. You should see the terminal that is running inside the docker container on the proxy server.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/675/1*ScmXQnhm-0vwQtGerSemZw.png" /></figure><p>Now we need to setup the web-terminal a little bit to get what we want.</p><p><strong>Setting up web-terminal</strong></p><ul><li>Do an apt update on the terminal.</li><li>apt install nano openssh-server to necessary packages.</li><li>Now generate the key-pairs using:<br>ssh-keygen -t rsa -C &quot;&lt;your email address&gt;&quot;</li><li>To enable one way SSH tunnel to the <em>target servers</em> using the <em>proxy server</em> as a medium, you have to add a config file inside the .ssh folder.</li><li>Open the config file in edit mode:nano ~/.ssh/config</li><li>Now add the following text and press Ctrl+X and then y to save the file. <em>Set User of the proxy server and change the IP addresses below:</em></li></ul><pre>Host proxy                    # A name for the proxy server<br>Hostname &lt;PROXY_SERVER_IP&gt;    # Provide the IP address of the proxy server<br>User root                     # The user to connect to in proxy server<br><br>Host targetA                  # A name to access to the Target Server A<br>ProxyCommand ssh -q proxy nc -q0 &lt;TARGET_A_SERVER_IP&gt; 22<br><br>Host targetB                  # A name to access to the Target Server B<br>ProxyCommand ssh -q proxy nc -q0 &lt;TARGET_B_SERVER_IP&gt; 22</pre><ul><li>Now access the target computer using: ssh &lt;target_a_username&gt;@targetA</li><li>To access <em>Target B </em>: ssh &lt;target_b_username&gt;@targetB</li></ul><h4>Security</h4><p>The web-terminal can be accessed from anywhere. How can it be secure? Well, every docker container is isolated from the host computer. So the web-terminal has no access to anything from the host-server. Even if you use the host server as the proxy, you will have to enter the password to access it via SSH.</p><p>There is only one tiny problem. That is anyone with the IP and port address of you terminal have access to it. You might not want to give access to even the <em>dockerized</em> terminal to just anybody.</p><p>Fortunately there is a way. You just have to pass the credentials when running docker container like this:</p><blockquote><em>docker run -itd --rm -p 7681:7681 tsl0922/ttyd ttyd -c &quot;&lt;username&gt;:&lt;password&gt;&quot; bash</em></blockquote><p>Now you will be prompt to enter the username and password whenever you access the web terminal.</p><h4>Why use Web Terminal?</h4><p>It is entirely possible to setup the client computers to use SSH tunnel directly, without using a web terminal. But there are several benefits in using a web terminal, e.g.:</p><ul><li>It gives you instant access. No need to setup every client computers!</li><li>You can access from any PC with an internet connection and a Web Browser. Imagine, when you need to update something quickly in the server from a friends PC.</li><li>It can be managed by an authority to add new servers or remove obsolete ones. You do not have to manually reconfigure SSH to all client computers whenever there are changes in the intranet.</li><li><a href="https://askubuntu.com/a/311454/457551">https://askubuntu.com/a/311454/457551</a></li><li><a href="https://unix.stackexchange.com/q/116191/257670">https://unix.stackexchange.com/q/116191/257670</a></li><li><a href="https://github.com/tsl0922/ttyd">https://github.com/tsl0922/ttyd</a></li><li><a href="https://hub.docker.com/r/tsl0922/ttyd/dockerfile">https://hub.docker.com/r/tsl0922/ttyd/dockerfile</a></li></ul><blockquote>Don’t forget to clap if you find this article useful!</blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=aee9b4c2c25e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Calculating the value of PI(π) from basic geometry (and some C++)]]></title>
            <link>https://medium.com/@dipu.sudipta/calculating-the-value-of-pi-%CF%80-from-basic-geometry-and-some-c-549e278afd03?source=rss-4135668b097------2</link>
            <guid isPermaLink="false">https://medium.com/p/549e278afd03</guid>
            <category><![CDATA[c-programming]]></category>
            <category><![CDATA[mathematics]]></category>
            <dc:creator><![CDATA[Sudipto Chandra]]></dc:creator>
            <pubDate>Sat, 29 Dec 2018 10:51:14 GMT</pubDate>
            <atom:updated>2018-12-29T11:26:56.900Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Jp862Bmjyf2o9Hjarbu_Gg.jpeg" /></figure><h3>Calculating the value of PI(<strong>π) using basic geometry </strong>(and some C++)</h3><p><em>In this article, I am going to explain how you can easily calculate the value of π using only basic knowledge of geometry.</em></p><h3>Base</h3><p>The area of a circle is <em>πr²</em>. If radius, <em>r = 1</em>, the area is only π. We just have to measure the area of a circle of radius 1 to find out the value of π. As we can not find the exact area directly, we will try to calculate an approximate value.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/251/1*DyWMP9pZ_iSaWikfYAq9hw.jpeg" /><figcaption>Figure 1: Rectangle <strong><em>ABCD </em></strong>has Area = 2</figcaption></figure><p>In the above circle, radius, <strong><em>OA </em></strong><em>= 1</em> and <strong><em>OB </em></strong><em>= 1</em>. Therefore,</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/288/1*phATlDBDch1aTIi_cTY9SQ.png" /></figure><p>The area of rectangle <strong>ABCD<em> </em></strong><em>= 2</em>. This number is our <strong>base</strong>. This is the initial approximation about the area of the circle.</p><h3>First parts</h3><p>Next we calculate the area of the remaining four parts of the circle outside the inner rectangle. All four parts are essentially similar. We can choose any one of them for our calculation.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/251/1*1FqJz4JHefE8rVTdR4W2Lw.jpeg" /><figcaption>Figure 2: There are 2² = 4 parts similar to <strong>ABC. </strong>Area of <strong>△ABC </strong>is (<strong><em>b ∙ h)</em></strong></figcaption></figure><p>The approximate area of the part <strong>ABC </strong>is the area of triangle <strong>△ABC</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/315/1*snuIc3PK61hB-UDBujLIhg.png" /></figure><p>As show in the figure 2, the base of base of <strong>△ABC </strong>= <strong>AB</strong>. As the line <strong>OC </strong>divides <strong>AB </strong>into two equal parts, half of the base,</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/81/1*wwAr6fL886iWXq3nEnVIgw.png" /></figure><p><strong>OC </strong>and and <strong>AB </strong>are perpendicular. The height,</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/356/1*L6yN7QOisK2z2SWKWUaIeQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/354/1*V4N-523gUmNCyGReF9XnmA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/396/1*hUTqJJe0ihfQp7U6-PYh6Q.png" /></figure><p>Therefore, the area of <strong>△ABC </strong>= <em>bh = </em>0.207106781187…</p><p>This is the area of a single part. Total area of four parts = 0.828427124746…</p><p>Adding this with the base, we get our second approximation of π= 2.828427124746…</p><h3>Second parts</h3><p>We are going to use the values of <em>b </em>and <em>h</em> we got from the first part here.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/504/1*MTRIShTAQSijxhuHEmR_bg.jpeg" /><figcaption>Figure 3: There are 2³ = 8 parts similar to <strong>AEC. </strong>Area of <strong>△AEC </strong>is (<strong><em>b2 ∙ h2)</em></strong></figcaption></figure><p>We already have the area of the <em>Octagon </em>inside the circle. It is our current approximation. Next we will find out the area of part <strong>AEC.</strong></p><p>There are total 2³ = 8 parts similar to this one. So we will multiply the area by 2³ after we find out the approximate area.</p><p>Here,</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/169/1*UAXV5pD5TRaG5qIyk8PjTw.png" /></figure><p>Next base,</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/169/1*FWkRsy1qzKYY7PrejtZ1Ww.png" /></figure><p>Next height,</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/200/1*yTymc1mGijGkU7q9Esn8kg.png" /></figure><p>We can easily calculate the next area now, which is:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/277/1*2LOOz-1fF0FlBLZjefnf2A.png" /></figure><p>Multiplying it by 2³ = 8 (as there are 8 similar parts), we get = 0.059977693337…</p><p>Adding it with our previous approximation of π we get = 3.061467458921…</p><p><em>We are getting close!</em></p><h3>Third parts and so one…</h3><p>Next up, we will get 2⁴ = 16 similar parts and have to use the second base and second height to get the third one.</p><p>But as you can guess this will take a lot of time to calculate third, forth and so on values by hand. I will cut it short and use computer for this repetitive job.</p><p>Here is a little <em>C++</em> program that will calculate the value of π using the above method:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/183002ff657f90ed3ba64f3878a0b4bd/href">https://medium.com/media/183002ff657f90ed3ba64f3878a0b4bd/href</a></iframe><p>Here are the first few approximate areas calculated by this program:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/041bbc42d4d6deae0d14a05c0cdac921/href">https://medium.com/media/041bbc42d4d6deae0d14a05c0cdac921/href</a></iframe><h3>Furthermore…</h3><p>Now, time for some exercises!</p><ul><li>Find a formula that does not depend on the previous base and height.</li><li>Can you minimize the expression of your formula?</li><li>Write a program that will give you the value of π up to whatever precision you want.</li></ul><p>I will continue discussing this topic in my next articles. The links will appear here after I have published them.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=549e278afd03" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[jquery.bangla — An input tool for bangla language]]></title>
            <link>https://medium.com/@dipu.sudipta/jquery-bangla-an-input-tool-for-bangla-language-cc3bd781b2d8?source=rss-4135668b097------2</link>
            <guid isPermaLink="false">https://medium.com/p/cc3bd781b2d8</guid>
            <category><![CDATA[javascript]]></category>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[npm]]></category>
            <category><![CDATA[jquery]]></category>
            <dc:creator><![CDATA[Sudipto Chandra]]></dc:creator>
            <pubDate>Thu, 02 Aug 2018 16:20:01 GMT</pubDate>
            <atom:updated>2018-08-06T06:51:24.490Z</atom:updated>
            <content:encoded><![CDATA[<p>It is quite vexing how little amount of resources are available for bangla language. There are only a couple of tools that allow users to type in bangla, and none are flexible enough to integrate into any type of web application.In the end, I decided to make an effort to build an input tool by myself.</p><blockquote>In this article, I am introducing a new typing tool designed for bangla language: <em>📦 </em><strong><em>jquery.bangla</em></strong><em> (</em><a href="https://github.com/dipu-bd/jquery.bangla"><em>Github</em></a><em>, </em><a href="https://npmjs.com/package/jquery.bangla"><em>NPM</em></a><em>)</em></blockquote><p>I was greatly inspired by this awesome tool: <a href="https://avro.im/">https://avro.im</a> to build this one. I used their <em>avro-phonetic</em> library in my package to display bangla word suggestions.</p><h3>💡 Live demo</h3><p>You can go visit and test this new tool right away at this link: <a href="https://dipu-bd.github.io/jquery.bangla/">https://dipu-bd.github.io/jquery.bangla</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/707/1*4ggK5t5unpje2f9n5WvmWg.png" /><figcaption>Live demo is available at <a href="https://dipu-bd.github.io/jquery.bangla/">https://dipu-bd.github.io/jquery.bangla/</a></figcaption></figure><h3>🔱 Features</h3><ol><li>This typing tools works well in both input, text-area and as well as content-editable divs. If you are using any WSYWYG html editors like: <a href="https://summernote.org/">summernote</a> or <a href="https://www.froala.com/wysiwyg-editor">froala</a>, you are good to go!</li><li>This input tools comes with a <em>built in suggestion provider</em>, that does not depend on any external libraries. It is especially made for bangla typing. With the help jquery’s <em>superpower</em>, it sits perfectly well with the editor.</li><li>In the backend, it uses <a href="http://torifat.github.io/jsAvroPhonetic/">avro-phonetic</a> with a large dictionary of bangla words to make suggestions. It saves user’s preferred words to give customized suggestions.</li><li>Works <em>offline</em>. It does not need any server to display word suggestions.</li><li>It keeps input <em>away</em> from the editor until user has confirmed it by using space, enter, tab or any other keys.</li><li>Initialization is <em>super easy</em>. Grab an element using jquery and use the .bangla function like this: $(&#39;some-input-element&#39;).bangla(configs);</li><li>You can use interact with it <em>easily</em> after initialization. e.g. You can toggle bangla typing mode by$(&#39;#input-area&#39;).bangla(&#39;toggle&#39;);</li><li>It is available in the <a href="https://npmjs.com/package/jquery.bangla">npm repository</a>.</li><li>I plan to keep up the support for this tool <em>as long as possible</em>.</li></ol><h3>🏥 <strong>Tutorial</strong></h3><p>This section is a brief guide about how to integrate jquery.bangla in you application.</p><h4>⮯ Installation</h4><p>There are several ways to install this tool into your website.</p><p><strong>First way (</strong>🖧<strong>CDN):</strong> You can integrate it directly into your html. Add the following script tags:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/71ea8403bfe3388e00915e4d57923785/href">https://medium.com/media/71ea8403bfe3388e00915e4d57923785/href</a></iframe><blockquote>Or, you can download the source files locally, and use them.</blockquote><p><strong>Second way (</strong>📦 <strong>NPM)</strong>: If you are using nodejs as you server, you can add <em>jquery.bangla</em> with your package manager as a dependency:</p><pre>npm i --save jquery.bangla jquery</pre><p>Later, in your javascript files, you can use either import or require to integrate this package into your website.</p><pre>import jquery from &#39;jquery&#39;;<br>window.$ = window.jQuery = jquery;</pre><pre>import &#39;jquery.bangla&#39;</pre><p>Or, if you do not have ES6 enabled, use the plain old require:</p><pre>window.$ = window.jQuery = require(&#39;jquery&#39;);</pre><pre>require(&#39;jquery.bangla&#39;);</pre><blockquote>The second way is particularly useful if you are using front-end frameworks like <a href="https://vuejs.org/">Vue</a>, <a href="https://reactjs.org/">React</a> or <a href="https://angular.io/">Angular</a>.</blockquote><h4>😊 Usage</h4><p>The input tool works in three different kinds of elements: <a href="https://developer.mozilla.org/en-US/docs/Web/HTML/Element/input">input</a> <a href="https://developer.mozilla.org/en-US/docs/Web/HTML/Element/textarea">textarea</a> and <a href="https://developer.mozilla.org/en-US/docs/Web/Guide/HTML/Editable_content">contenteditable divs</a> (In fact, any type of element that has keydown events and selection range should be okay, but I have not tested them).</p><p>Suppose, you have a textarea in your web-page like this one:</p><pre>&lt;textarea id=&quot;text-area&quot; placeholder=&quot;Start typing in bangla&quot;&gt;&lt;/textarea&gt;</pre><p>To enable bangla typing inside this element, you have to initialize it first:</p><pre>$(&#39;#text-area&#39;).bangla({ enable: true });</pre><p>The function .bangla() accepts a config object, but it will also work without any parameters. You will find the list of available configs <a href="https://github.com/dipu-bd/jquery.bangla/blob/master/README.md">here</a>.</p><p>If you wish to <em>toggle</em>, <em>disable</em>, or <em>enable</em> bangla typing you can pass the following parameters:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/a3b34325c96ac521ca0c465e9eb87ed4/href">https://medium.com/media/a3b34325c96ac521ca0c465e9eb87ed4/href</a></iframe><h3>🍀 Thank you</h3><p>The package is published under <a href="https://github.com/dipu-bd/jquery.bangla/blob/master/LICENSE">Mozilla Public License 2.0</a>. You are free to modify and use it as long as you follow the guidelines.</p><p>Thank you for reading this article. Your voice of encouragement and support means a lot to me. So do not hesitate to put a 🟊 on <a href="https://github.com/dipu-bd/jquery.bangla">this repository</a> and report any issues you see to help it keep growing.</p><p>👏 <strong><em>PS</em></strong><em>: If you like article do not forget to give some claps. </em>👏</p><ul><li>Main repository: <a href="https://github.com/dipu-bd/jquery.bangla">https://github.com/dipu-bd/jquery.bangla</a></li><li>Live demo: <a href="https://dipu-bd.github.io/jquery.bangla">https://dipu-bd.github.io/jquery.bangla</a></li><li>NPM package: <a href="https://npmjs.com/package/jquery.bangla">https://npmjs.com/package/jquery.bangla</a></li><li>Example: <a href="https://github.com/dipu-bd/jquery.bangla/tree/gh-pages">https://github.com/dipu-bd/jquery.bangla/tree/gh-pages</a></li><li>Inspiration: <a href="https://github.com/omicronlab/avro-pad">https://github.com/omicronlab/avro-pad</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cc3bd781b2d8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Savior of the day: <no-ssr>]]></title>
            <link>https://medium.com/@dipu.sudipta/savior-of-the-day-no-ssr-4235f57268eb?source=rss-4135668b097------2</link>
            <guid isPermaLink="false">https://medium.com/p/4235f57268eb</guid>
            <category><![CDATA[vuejs]]></category>
            <category><![CDATA[nuxtjs]]></category>
            <category><![CDATA[javascript]]></category>
            <dc:creator><![CDATA[Sudipto Chandra]]></dc:creator>
            <pubDate>Fri, 08 Jun 2018 08:38:05 GMT</pubDate>
            <atom:updated>2018-07-31T10:08:32.069Z</atom:updated>
            <content:encoded><![CDATA[<p>A few days ago, I run into a serious memory leak issue with <a href="https://dungulie.com/">my company website</a> I am maintaining. I want to share this story in case someone faces a similar problem.</p><p>I have applied Vue in several small to medium scale apps beforehand, but never tried <a href="https://nuxtjs.org/">Nuxt</a> or SSR (Server Side Rendering). Let me remind you of a particular line in the <a href="https://vuejs.org/v2/guide/computed.html#Computed-Caching-vs-Methods">vue.js guide</a>:</p><blockquote>computed properties are cached based on their dependencies.</blockquote><p>I failed to understand the impact of it will have on SSR in time. As you may know, the SSR only call <em>beforeCreate()</em> and <em>create() </em>functions. But <em>destroyed()</em> or <em>beforeDestroy()</em> methods are left for the browser only. As a result, we do not have a chance to clear some cached values in server side.</p><p>In my web app, I needed to initialize a large list (of about <em>150 KB</em> in size) every time a new request came for a page. On top of it, each list is different for different users. This list is fetched internally inside of the server in SSR. And some computed properties makes some calculation and clone some parts of it. For the sub-components, more computed property was in use. In total, it was a huge chunk of computed data being calculated and cached for each client request!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*d-ZbdImPx4zRW0zK4QL49w.jpeg" /></figure><p>When I tested the server, it could not serve more than 100 concurrent requests before freezing up. Well, it might not be a problem with Vue or SSR, but it can be a limitation of my knowledge and <em>bad designing</em>, but I still do not see any way out of this without throwing out SSR or writing separate code for client and server.</p><p>The temporary solution to this issue I came up with is to block the SSR for some special components. Nuxt supports a component called <a href="https://nuxtjs.org/api/components-no-ssr/"><strong>&lt;no-ssr&gt;</strong></a>. The server side rendering engine ignores everything inside of it. I used it everywhere I could, exposing only a few lines of gentle codes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1008/1*K-Dk3UD0D3F__CG-AGREXg.png" /><figcaption>Use <strong>&lt;no-ssr&gt;</strong> to disable server-side-rendering of some components</figcaption></figure><p>Although this can not be said to be an actual solution, but it bypasses the issue. I needed only a few things to come from server, like- meta data, and some small components which take little but negligible memory and processing power.</p><p>The real solution would be to change the architecture of entire web app in a way that produce very little memory footprint before mounting it in the browser. But isn’t it too much troublesome?!</p><p><em>Mommy… I do not want SSR…</em> <em>Can I just use simple template rendering engines, like: ejs, pug… please!</em></p><p>Before you pick up SSR, or do anything rush, make sure to read these nice articles to see if you have another workaround of your problem:</p><ul><li><a href="https://codeburst.io/when-to-use-server-side-rendering-ssr-in-vue-js-projects-697bd925d57b">https://codeburst.io/when-to-use-server-side-rendering-ssr-in-vue-js-projects-697bd925d57b</a></li><li><a href="https://medium.com/js-dojo/faking-server-side-rendering-with-vue-js-and-laravel-65f7fe88aea8">https://medium.com/js-dojo/faking-server-side-rendering-with-vue-js-and-laravel-65f7fe88aea8</a></li><li><a href="https://medium.com/vue-mastery/best-practices-for-nuxt-js-seo-32399c49b2e5">https://medium.com/vue-mastery/best-practices-for-nuxt-js-seo-32399c49b2e5</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4235f57268eb" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>