Re: [PATCH 07/17] cryptodisk,luks: Allow special processing for comparing UUIDs.

[Glenn Washburn]

On Thu, 30 Jul 2020 17:24:12 +0200
Patrick Steinhardt <ps@pks.im> wrote:

> On Wed, Jul 29, 2020 at 04:50:12PM -0500, development@efficientek.com
> wrote:
> > From: Glenn Washburn <development@efficientek.com>
> > 
> > Create grub_uuidcasecmp to compare UUIDs in a case-insensitive
> > manner and that ignores '-' characters. This is backwards
> > compatible with the old LUKS1 code that stored and compared against
> > UUIDs without dashes. However, the new LUKS2 code stores and
> > compares UUIDs that contain dashes. Really, the UUID comparison
> > shouldn't care about the dashes, as this change implements. Now
> > your old scripts will continue to work with UUIDs without dashes,
> > but you may choose to use UUIDs with dashes now too for both LUKS1
> > and LUKS2.
> 
> FYI, this is crossing with my own patchset to implement probing
> support for LUKS2 [1]. I do like that we start becoming agnostic of
> those dashes now, though, as it really is kind of annoying to handle
> as a user.
> 
> 1: https://lists.gnu.org/archive/html/grub-devel/2020-05/msg00235.html

Yep, I saw your patch and like that it brings into alignment LUKS1 and
LUKS2 UUID handling.  Unfortunately, LUKS UUID handling is also
different than FS UUID handling, which uses dashes.  So not only is the
current behavior annoying, its also confusing.  I've been bitten by this
a couple times.  This change brings LUKS* UUID handling into alignment
with FS UUID handling and is backwards compatible with current LUKS
handling.  Since your patch set hasn't been committed, would you be
opposed to dropping that patch?