On Sun, Jun 20, 2021 at 10:39:13AM -0400, Theodore Ts'o wrote: > On Sat, Jun 19, 2021 at 08:44:52PM -0700, Josh Triplett wrote: > > On Thu, Jun 17, 2021 at 10:36:48AM +0200, Paul Gevers wrote: > > > On Tue, 08 Jun 2021 20:22:39 -0700 Josh Triplett <j...@joshtriplett.org> > > > wrote: > > > > Package: e2fsprogs > > > > Version: 1.46.2-2 > > > > Severity: important > > > > X-Debbugs-Cc: j...@joshtriplett.org > > > > > > > > "important" because this does cause data loss; not filing as > > > > release-critical because using mke2fs on a full disk or full disk image > > > > with a specified size to write a partition may be a niche use case. > > > > > > Do you know if this is a regression with respect to the version in > > > bullseye (1.46.2-1)? > > > > I'm fairly sure this bug has existed for quite a long time. Please don't > > let it affect propagation of new versions. > > Indeed, this bug has been around since 2008, and most of the time, > people only use the size option when they are creating an file system > in a image file, where this wouldn't matter. And if they are running > mke2fs on a partition with a size limit, which is unlikely to begin > with, it's even more unlikely they would care about what came after > the file system in the partition. > > That's why no one noticed or complained for 13 years; and so I would > consider "important" to be a stretch. :-)
The only reason I used "important" was because it could cause data loss; I *absolutely* agree that it's an obscure combination of options, though one that seemed natural to use together. I only figured out this issue after spending a while debugging why a virtual machine image wouldn't boot, before finding that the subsequent EFI system partition had been partially overwritten. - Josh
