Re: GSoC 2026: Interest in ext2fs and systems programming

On Mon Mar 2, 2026 at 9:12 PM CET, Matthias Andree wrote:
>
> What concretely do you want to do about the "ext2fs user space tools"? 

My bad for not being specific, I meant a project listed under that title
in the list of project ideas for GSoC. 

> The extant fs relevant tools are brought over from the Linux world, are 
> pretty mature, and are certainly not a suitable object for a summer 
> project, and unless the kernel performance of ext4fs is up to speed 
> (meaning performance), why would you want to invest time on a tool chain 
> that is mature but isn't FreeBSD's?

The current driver lacks some features, and the project would mean
implementing them so that they ship natively, instead of having to rely
on third-party, GPL software. Including it in FreeBSD would also make
installation smoother.

I believe the technical scope is significant. It means starting from the
NetBSD code that provides support for ext2, and augmenting it to support
ext3 and ext4, a considerable algorithmic challenge that would give
FreeBSD native autonomy with these filesystems.

I understand the weaknesses of this idea, but the advantages are also
substantial. If the community finds it not worth the time, I'll be glad 
to switch my focus to another project.