This is a useful addition, thank you!

Ideally, a reader interface would also be great, but this is much more complicated.

Ideally, a reader interface would also be great, but this is much more complicated.

If i'm motivated i'll take a look into it. I think if we base ourserves on Reader's readBoundedBytes we should be able to chunk the decoding there too?

I realize i removed the comptime keyword on a test that was explicitely comptime. I should try to find a way to avoid this.

There, using BoundedArray the tests works at comptime, while still showing that it uses Writer's writeAll.

Ideally, a reader interface would also be great, but this is much more complicated.

So i looked into it for a bit, and it's actually not that hard. The Reader interface is pretty neat and allows us to do this quite easily, so i added this interface in ab740ca

This means in theory that you could encode a file directly to base64 without having to load it in memory. Sounds pretty nice !

while i'm at it, should i try to do the reverse for decoding ?

while i'm at it, should i try to do the reverse for decoding ?

That could be great, if only for consistency. But in a distinct PR.

Is there anything missing for this PR @jedisct1 ? Not trying to stress you on the matter, genuinely asking, as I do not know if there are any other prerequisites.

Thank you for your feedback.

@marler8997 i applied your suggestions to the other method in b9dfe0f

Now with a zig fmt pass that was missing, sorry.

Hey there. Anything missing here so it can be merged? I'm planning on doing the decoding part soon in another PR.

thanks @Vexu for the auto-merge. Could you reenable it? I just rebased again. There was an error in aarch-linux-release, but it wasn't related to my stuff.