How to handle very large zip files on browser

[mbret]

I am the developer of https://oboku.me/ and I am trying to provide support for large books. Some comics can be up to a GB size and I have two issues currently. Some context first:

• I download the files on the browser as blob
• Once downloaded, the blob is stored in indexeddb
• the blob is later loaded in jszip to be displayed.

Currently, the download crash on devices that obviously don't have enough memory. Loading large books on chrome is also extremely slow due to jszip I would guess.

So I am trying to solve two issues basically.

• Download large files (larger than RAM)
• Load large files (larger than RAM).

First is that even doable on a browser ?
Second, I thought of streaming the download directly to the File System API but safari does not support stream and it is the browser which with I have most issue so for the time being it seems indexeddb is more reasonable. I thought about streaming the download and storing chunks in indexeddb but before doing so I wanted to make sure I would be able to consume the chunks with fflate in a way that would work and also not saturate the RAM ?

Well, any insight on how to solve my issue would be appreciated.

(side note: I am looking for a full browser experience, no server involved for streaming allowed)

[Laurens256]

I'm using native-file-system-adapter for this. Take a look at fileHandle.createWritable() and how to set up a backup using a service worker