Yes. In fact, I believe it already has. Kernels in the development 2.1.x
series have not zeroed indirect blocks since more than six months ago. At the
beginning of December 1996, there was some talk on the linux-kernel
mailing-list of producing another 2.0.x production kernel that also leaves
indirect blocks intact on deletion. Although as of the pre-released versions
of kernel 2.0.31 this has not happened, I suspect that it is feasible. Once
Linus and the other kernel hackers overcome this limitation in the production
kernels, a lot of my objections to the technique of modifying inodes by hand
will disappear. At the very latest, this should happen on the release of the
2.2.x kernel series, which (according to historical kernel development
time-scales) should happen some time in the first quarter of 1998. When this
wart is corrected, it will also be possible to use the dump command in
debugfs on long files.