The ABI status of filesystem formats

I understand the desire to use ext4 for a read-only fs as a need for features maybe not-supported in certain other filesystems.

I don't understand a reluctance to certify e2fsck as the arbiter of what may be mounted as an ext2/3/4 filesystem. If it is discovered that some bit pattern is a security hole, surely e2fsck will be updated to patch up an image that tickles that security hole?

It seems not necessary to remain bitwise compatible with filesystems that *current* e2fsck would alter, but that older e2fsck would allow. e2fsck should continue to accept non-broken images, and to fix known-broken ones. So, we could accept a future kernel requiring that a file system image that used to mount as-is be fixed by a new e2fsck before it may be mounted. We could even have e2fsck refuse to fix certain old images it finds too confusing, as it might for any damaged image.