why dont you use a stream to generate file?

I build a service which can generate arbitrary size file on Cloudflare Workers using stream.

https://github.com/alsotang/cf_workers__file

I wanna copy your implement at first, but it seems you alloc a corresponding size of memory to hold the file. Your implement would reach the worker's memory limit if the file size is big.

Start by locating the current file-generation entry point in this repository, then compare its behavior with the stream-based example at https://github.com/alsotang/cf_workers__file. Determine whether large generated files can be produced without allocating the full size in memory. Done means the implementation avoids exceeding Cloudflare Workers memory limits for large files and existing behavior remains intact.

Written by the indexing model from the issue text.