Re: [PATCH 3/3] erofs-utils: tests: check that the xattr layout is order-independent

[Gao Xiang <xiang@kernel.org>]

Hi Martin,

On Wed, Aug 26, 2026 at 09:18:47PM +0200, Martin Pitt wrote:
> Hello again,
> 
> Martin Pitt [2026-08-04  5:32 +0200]:
> > b260119b8edb ("erofs-utils: mkfs: emit an inode's xattrs in a canonical
> > order") made the layout independent of the order listxattr(2) reports an
> > inode's attributes in. Cover it: stage the same tree twice with the same
> > attributes set in opposite order, and compare the two images. The names
> > differ in length and share prefixes, so a subtly wrong comparison cannot
> > order them correctly by accident.
> 
> I see my first [1] and second [2] preparatory patch made it into the tree
> [3][4], but not this third one (and the main test).
> 
> Do you want me to explain or change anything, or did it simply fell through the
> cracks?

I didn't miss this one, there are also other pending tests in my queue,
I need to resolve them all but maybe a bit delay.

It's already merged with very minor changes.

Thanks,
Gao Xiang

> 
> Thank you!
> 
> Martin
> 
> [1] https://lore.kernel.org/linux-erofs/20260804033214.211267-2-martin@amutable.com/
> [2] https://lore.kernel.org/linux-erofs/20260804033214.211267-3-martin@amutable.com/
> [3] https://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs-utils.git/commit/?h=experimental-tests&id=4b02ea1792c14cb9f250bd55aced5614e699397a
> [4] https://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs-utils.git/commit/?h=experimental-tests&id=887ce207bc8b068de9771b736333a8da6d939670
>