Turbopack: switch from base40 to base38 hash encoding (remove ~ and . from charset)#91832
Merged
Conversation
Contributor
Failing test suitesCommit: 26bf3a1 | About building and testing Next.js
Expand output● next-config-ts-tsconfig-extends-esm › should support tsconfig extends (ESM)
Expand output● nx-handling › should work for pages page ● nx-handling › should work for pages API ● nx-handling › should work with app page ● nx-handling › should work with app route ● Test suite failed to run |
Merging this PR will improve performance by 4.51%
Performance Changes
Comparing 
Footnotes
|
Contributor
Stats from current PR✅ No significant changes detected📊 All Metrics📖 Metrics GlossaryDev Server Metrics:
Build Metrics:
Change Thresholds:
⚡ Dev Server
📦 Dev Server (Webpack) (Legacy)📦 Dev Server (Webpack)
⚡ Production Builds
📦 Production Builds (Webpack) (Legacy)📦 Production Builds (Webpack)
📦 Bundle SizesBundle Sizes⚡ TurbopackClient Main Bundles
Server Middleware
Build DetailsBuild Manifests
📦 WebpackClient Main Bundles
Polyfills
Pages
Server Edge SSR
Middleware
Build DetailsBuild Manifests
Build Cache
🔄 Shared (bundler-independent)Runtimes
📎 Tarball URL |
styfle
reviewed
Mar 24, 2026
…harset The `~` and `.` characters in base40-encoded filenames (e.g. `turbopack-0c3o1svijj_~~.js`, `0...f7~att2_2.js`) are blocked by Nginx hardening rules and enterprise WAF configurations, causing 403 Forbidden errors on deployment. Remove these characters to produce filenames that are safe for all common web server configurations. Fixes #91678 Co-Authored-By: Claude <noreply@anthropic.com>
Co-Authored-By: Claude <noreply@anthropic.com>
Co-Authored-By: Claude <noreply@anthropic.com>
Regenerated all snapshot output files using UPDATE=1 to reflect the new base38 hash encoding (charset `0-9a-z_-`, dropping `~` and `.`). Co-Authored-By: Claude <noreply@anthropic.com>
Static asset filenames embedded in the inline snapshot tests now use base38-encoded hashes instead of base40. Co-Authored-By: Claude <noreply@anthropic.com>
Co-Authored-By: Claude <noreply@anthropic.com>
241d01b to
26bf3a1
Compare
styfle
reviewed
Mar 24, 2026
| /// information loss. | ||
| pub const BASE40_LEN_64: usize = digits_for_bits(BASE, 64); | ||
| pub const BASE38_LEN_64: usize = digits_for_bits(BASE, 64); |
Member
There was a problem hiding this comment.
In my opinion we should just increase from 64bit (13 chars) to 84bit (16 chars).
That would be an increase of 252x meaning 751 billion IDs until a 1% chance of collision.
Then we could probably omit the 128bit manifest.
styfle
approved these changes
Mar 24, 2026
Contributor
|
so we avoid uppercase characters due to case insensitive file systems, but URLs are case sensitive, so using upper case character is fine we just don't want to produce two outputs that differ only by case. This is essentially a hash collision... do we need to care about it? |
lukesandberg
approved these changes
Mar 24, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What?
Fixes a regression from #91137
Switch Turbopack's hash encoding charset from base40 (
0-9 a-z _ - ~ .) to base38 (0-9 a-z _ -), removing the~and.characters. Pure rename/charset change — no structural changes.Why?
The
~and.characters in base40-encoded filenames are blocked by standard Nginx hardening rules (block_common_exploits.conf) and enterprise WAF configurations, causing 403 Forbidden errors when applications are deployed behind security-hardened infrastructure.Examples of problematic filenames:
turbopack-0c3o1svijj_~~.js—~~flagged as directory traversal / injection0...f7~att2_2.js—...flagged as path traversal attempt0q~2copru0zy0.css—~filtered by some WAF rulesetsPrevious hex-only filenames (e.g.
turbopack-01ca012029ca2e66.js) had no such issues.Fixes #91678
How?
Charset change (
turbo-tasks-hash/src/base38.rs, renamed frombase40.rs):0123456789abcdefghijklmnopqrstuvwxyz_-BASE40_*→BASE38_*,encode_base40→encode_base3838^13 > 2^64,38^25 > 2^128)HashAlgorithmenum variants (turbo-tasks-hash/src/lib.rs):Xxh3Hash64Base40→Xxh3Hash64Base38Xxh3Hash128Base40→Xxh3Hash128Base38Bit computation comment (
turbopack-core/src/ident.rs):7 base38 chars ≈ 37 bits→≈ 36 bits(log2(38) × 7 = 36.7)Test regex patterns (15 test files):
[0-9a-z_.~-]→[0-9a-z_-]to match the new charset