How-to

Website

Web app

Check a deploy

This recipe is the list to go through after each deploy, on any host: the mistakes it catches pass every check on your own machine. You need the deployed site's address. - Open a deep link directly and reload it. A 404 means the host does not fall back to index.html (a single-page build), or base_path is wrong. - View a page's source. In a static build, the page's text is in the HTML. - Check the response headers in the browser's network panel. - Run a link-preview debugger on a page, and submit sitemap.xml to the search engines' consoles. The same checks from a terminal, with your site's address in SITE and a page other than the home page in place of /docs:

bash

$ SITE=https://example.com $ curl -sI "$SITE/docs" | head -1 # HTTP/2 200, not 404 $ curl -s "$SITE/docs" | grep -o '<h1[^>]*>[^<]*' # a static build: the heading is in the HTML $ curl -sI "$SITE/" | grep -iE 'content-security-policy|x-frame-options|cache-control' $ curl -s "$SITE/sitemap.xml" | head # absolute addresses under meta.site_url

How it works

Each line prints something when the deploy is right: a 200 status, the page's <h1> and its text, the three headers, and the first entries of the sitemap. An empty line is the check that failed.

For the link preview, paste a page's address into the Facebook Sharing Debugger or LinkedIn's Post Inspector; for the sitemap, submit $SITE/sitemap.xml in Google Search Console and Bing Webmaster Tools.

Before the next deploy, wf verify loads every page of the build in headless Chrome and catches a script error, a failed request or a blank page before it ships (Install wf on a build machine).