4 ms·
The heuristics used to attempt to differentiate between a so-called "bot" and a "human" are, IMHO, inadequate as long as there are "humans" that are allegedly m
by 1vuio0pswjnm7 3y ago
The heuristics used to attempt to differentiate between a so-called "bot" and a "human" are, IMHO, inadequate as long as there are "humans" that are allegedly mistaken for "bots" and blocked. "Use Chrome" is not a solution. A person using Firefox or some other non-Google software is still a "human". But not according to these brilliant "site protection" schemes. What level of false positives is acceptable.
Using JS to "verify that this is not a bot" is a way to force users to enable JS and expose themselves to more advertising.
- sam0x17 3y agoFurthermore, all bots worth their salt as far as threats go enable js and do everything they can to appear like a normal browser.
- azalemeth 3y agoI'd love to know if puppeteer passed that test (probably). I have had exactly this problem many, many times and it is incredibly frustrating.
- 20after4 3y agopuppeteer passes the test if you run it from a machine that already has a good cloudflare reputation score. Try it from an AWS instance and it definitely fails 100% of the time. (I've tried it, that's how I know)
- sam0x17 3y agoThere are github projects that are forks of things like selenium and puppeteer that are specifically designed to avoid detection for things like scraping google search results, etc.
- tick_tock_tick 3y agoThat's fine as processing the javascript increases the coast at-least.
- account42 3y agoBlocking bots in the first place should not be acceptable since bots only act on behalf of humans. What should be blocked is abusive behavior that actually impacts the site - a single one off GET to what should be a static page should never be blocked, yet that's what CF does.