From 06be7f7567fa23d85619369b48eb62b056a60656 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Marcos=20C=C3=A1ceres?= <marcos@marcosc.com>
Date: Wed, 9 Jun 2021 18:32:43 +1000
Subject: [PATCH] fix(validate/links): use 'link-checker' temporarily (#78)

Co-authored-by: Sid Vishnoi <8426945+sidvishnoi@users.noreply.github.com>
---
 docs/options.md       |  2 ++
 src/validate-links.ts | 30 ++++++++++++++++++++----------
 2 files changed, 22 insertions(+), 10 deletions(-)

diff --git a/docs/options.md b/docs/options.md
index 492e0a0..98952de 100644
--- a/docs/options.md
+++ b/docs/options.md
@@ -102,6 +102,8 @@ The Action will try to make use of metadata/config from previously published ver
 
 Whether or not to check for broken hyperlinks.
 
+**Warning:** This feature is experimental.
+
 **Possible values:** true, false
 
 **Default:** false
diff --git a/src/validate-links.ts b/src/validate-links.ts
index c0f8144..a011de4 100644
--- a/src/validate-links.ts
+++ b/src/validate-links.ts
@@ -1,24 +1,34 @@
 import { env, exit, install, sh, yesOrNo } from "./utils.js";
-import { PUPPETEER_ENV } from "./constants.js";
-
 import { BuildResult } from "./build.js";
 type Input = Pick<BuildResult, "dest" | "file">;
 
+const URL_IGNORE = [
+	// Doesn't like robots
+	"https://ev.buaa.edu.cn/",
+	// The to-be published /TR URL.
+	// Ideally should include shortname, but may be good enough.
+	`/TR/.+${new Date().toISOString().slice(0, 10).replace(/-/g, "")}/$`,
+];
+
 if (module === require.main) {
 	if (yesOrNo(env("INPUTS_VALIDATE_LINKS")) === false) {
 		exit("Skipped", 0);
 	}
-	exit("Link validator is currently disabled due to some bugs.", 0);
 
 	const input: Input = JSON.parse(env("OUTPUTS_BUILD"));
 	main(input).catch(err => exit(err.message || "Failed", err.code));
 }
 
-export default async function main({ dest, file }: Input) {
-	await install(`href-checker`, PUPPETEER_ENV);
-	await sh(`href-checker ${file} --no-same-site`, {
-		output: "stream",
-		cwd: dest,
-		env: PUPPETEER_ENV,
-	});
+export default async function main({ dest: dir }: Input) {
+	await install("link-checker");
+	const opts = getLinkCheckerOptions(URL_IGNORE);
+	// Note: link-checker checks a directory, not a file.
+	await sh(`link-checker ${opts} ${dir}`, "stream");
+}
+
+function getLinkCheckerOptions(ignoreList: string[]) {
+	return ignoreList
+		.map(url => `--url-ignore="${url}"`)
+		.concat(["--http-timeout=50000", "--http-redirects=3", "--http-always-get"])
+		.join(" ");
 }