fix(netlify, vercel): order route rules from most specific + avoid double-rendering root #768
Conversation
src/presets/netlify.ts
Outdated
| nitro.options.routeRules | ||
| ).filter(([_, routeRules]) => routeRules.headers)) { | ||
| const rules = Object.entries(nitro.options.routeRules) | ||
| .sort((a, b) => b[0].split("/").length - a[0].split("/").length) |
There was a problem hiding this comment.
Small issue with sorting /test/a is more specific than /test/* but if * is before, it remains before and shadows.
There was a problem hiding this comment.
nice catch - thank you β€οΈ
src/presets/vercel.ts
Outdated
| src: "/(.*)", | ||
| dest: "/__nitro", | ||
| }, | ||
| ...(!nitro.options.routeRules['/**']?.cache || !(nitro.options.routeRules['/**'].cache.swr || nitro.options.routeRules['/**']?.cache.static) |
There was a problem hiding this comment.
Can you please explain it better in line above? Do we need to also handle other route rules for catch-all such as headers?
Codecov Report
@@ Coverage Diff @@
## main #768 +/- ##
==========================================
- Coverage 70.07% 70.07% -0.01%
==========================================
Files 57 57
Lines 5209 5229 +20
Branches 568 575 +7
==========================================
+ Hits 3650 3664 +14
- Misses 1550 1556 +6
Partials 9 9
Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. |
|
Thanks for the updates. About sorting, it should be same as radix3 structure while hoisting placeholders in regex, might be less accurate i think we can merge it to unblock downstream issue for nuxt.com :) |
π Linked issue
β Type of change
π Description
This orders route rules within vercel + netlify presets so that (particularly glob) rules do not shadow rules below them. In addition, if the default vercel route rule is
/**: { static: true }we don't also need to output a second, SSR rule.cc: @antfu - if you clone this and checkout commit c7f8f39, then run
vitest run -uthe test passes and the inline snapshots are not updated. I've also noticed some other weird issues like a different snapshot within a file being updated than the one it corresponds to (e.g. the next snapshot below).π Checklist