fix: sanitize some slugs that are React router dynamic routes

[Josh-Cena]

Motivation

Fix #5418. Not sure if this should be breaking change, since these paths probably can't be resolved previously anyways...

Have you read the Contributing Guidelines on pull requests?

Yes

Test Plan

Unit tests + dogfooding

[netlify]

✔️ [V2]

🔨 Explore the source changes: 40661d0

🔍 Inspect the deploy log: https://app.netlify.com/sites/docusaurus-2/deploys/61f61ab419a5ef0008d3337c

😎 Browse the preview: https://deploy-preview-6510--docusaurus-2.netlify.app

[github-actions]

⚡️ Lighthouse report for the changes in this PR:

Category	Score
🟠 Performance	72
🟢 Accessibility	100
🟢 Best practices	100
🟢 SEO	100
🟢 PWA	92

Lighthouse ran on https://deploy-preview-6510--docusaurus-2.netlify.app/

[github-actions]

Size Change: +3.37 kB (0%)

Total Size: 759 kB

Filename	Size	Change
website/.docusaurus/globalData.json	47.9 kB	+953 B (+2%)
website/build/assets/js/main.********.js	576 kB	+2.41 kB (0%)

ℹ️ View Unchanged

Filename	Size
website/build/assets/css/styles.********.css	105 kB
website/build/index.html	30.1 kB

_{compressed-size-action}

[netlify]

✔️ [V2]

🔨 Explore the source changes: f747849

🔍 Inspect the deploy log: https://app.netlify.com/sites/docusaurus-2/deploys/61f61c186e1f8a0008ce37d9

😎 Browse the preview: https://deploy-preview-6510--docusaurus-2.netlify.app

[slorber]

I have one concern with such change, because we should probably have distinct logic for 2 cases:

• slug inferred from the FS structure: yes we want to failsafe and sanitize those to compute a good default slug
• slug explicitly provided by the user: we should rather warn the user that some slugs are wrong and are sanitized

Are (parens) really illegal? Isn't there a way to escape special chars?

Afaik ReactRouter is using https://github.com/pillarjs/path-to-regexp

I'd like to understand how other SSGs handle those cases before taking a decision on what is best for Docusaurus

[Josh-Cena]

Are (parens) really illegal? Isn't there a way to escape special chars?

They are not illegal slugs by the URL spec, but React router interprets them as a regex capturing group. So if you have a page with slug /doc(1), it's actually served at /doc1

[slorber]

So if you have a page with slug /doc(1), it's actually served at /doc1

But is it the case with Gatsby / Next.js / Remix too? that's worth double-checking and try to conform to what they do, particularly if someday we want to use another SSG as the underlying framework behind Docusaruus