Fetching ehow.com times out, it can never be fetched, even by puppeteer. Still, guessed icon is a valid result. It's the only case out of top 500 where puppeteer crawling of index page fails.
Even when finding icons works, but downloading them e.g. with download, might fail. It's ok from this library's perspective, but favicon-thief users need to be aware of it. It fail for the following domains (5 cases out of 500):
We might want to extract getFavicons as a separate npm package. Then we'd move the (huge) puppeteer and node-fetch dependencies there so favicon-thief stays lightweight and fetch-agnostic
APIs changed drastically, so I will bump major package version
Favicons of all top 500 websites can now be discovered :tada:
probe-image-size
:needle
which uses native node http & https clients, which - likenode-fetch
- fail with multiple websites that have anti-bot protectionsizes
attribute is good enoughO(n)
toO(1)
node-fetch
as it is much faster thanpuppeteer
and works in great majority of cases (>95%)puppeteer
. Still, guessed icon is a valid result. It's the only case out of top 500 where puppeteer crawling of index page fails.favicon-thief
users need to be aware of it. It fail for the following domains (5 cases out of 500):getFavicons
as a separate npm package. Then we'd move the (huge)puppeteer
andnode-fetch
dependencies there sofavicon-thief
stays lightweight and fetch-agnostic