As a result, parse returns wrong cache for these files, which can cause builds to fail randomly or produce unexpected results. As mentioned in https://github.com/bevacqua/hash-sum/issues/4, hash-sum algorithm is not suitable for content hashing. Unfortunately, I cannot share these private files here, but this issue may arise in projects with more than 10,000 files.
Would it be possible to switch to a more promising hash function, such as xxhash?
Version
3.3.0
Explanation
I have found two different files that has the same hash value on following lines:
https://github.com/vuejs/component-compiler-utils/blob/27e4fa359d77d28b73a66089ce781e0a03962277/lib/parse.ts#L55-L57
As a result,
parse
returns wrong cache for these files, which can cause builds to fail randomly or produce unexpected results. As mentioned in https://github.com/bevacqua/hash-sum/issues/4,hash-sum
algorithm is not suitable for content hashing. Unfortunately, I cannot share these private files here, but this issue may arise in projects with more than 10,000 files. Would it be possible to switch to a more promising hash function, such asxxhash
?Related issue
https://github.com/vitejs/vite/issues/912