Optional case-insensitivity

Optional case-insensitivity

[Neal Gompa]

Hey,

So I saw today on LWN an article about ext4 gaining the ability to do
per file/folder case-insensitivity[1]. I can see some value in this
property existing for subvolume/folder/file level for Btrfs for things
like Wine and Darling, which could take advantage of this to help
support Windows and macOS applications that expect insensitivity to
work properly. In particular, I'm looking at how games are glitchy
with case sensitivity because it's rarely tested (both Windows and
macOS are case-insensitive by default).

Has anyone looked at what it would take to do this in Btrfs too?

[1]: https://lwn.net/Articles/829737/

-- 
真実はいつも一つ！/ Always, there's only one truth!

Re: Optional case-insensitivity

[Chris Murphy]

On Fri, Aug 28, 2020 at 1:44 PM Neal Gompa <ngompa13@gmail.com> wrote:
>
> Hey,
>
> So I saw today on LWN an article about ext4 gaining the ability to do
> per file/folder case-insensitivity[1]. I can see some value in this
> property existing for subvolume/folder/file level for Btrfs for things
> like Wine and Darling, which could take advantage of this to help
> support Windows and macOS applications that expect insensitivity to
> work properly. In particular, I'm looking at how games are glitchy
> with case sensitivity because it's rarely tested (both Windows and
> macOS are case-insensitive by default).
>
> Has anyone looked at what it would take to do this in Btrfs too?
>
> [1]: https://lwn.net/Articles/829737/

It might also make virtiofs on Btrfs more compelling across
virtualized Windows and macOS.


-- 
Chris Murphy

Re: Optional case-insensitivity

[Chris Murphy]

On Fri, Aug 28, 2020 at 5:26 PM Chris Murphy <lists@colorremedies.com> wrote:
>
> On Fri, Aug 28, 2020 at 1:44 PM Neal Gompa <ngompa13@gmail.com> wrote:
> >
> > Hey,
> >
> > So I saw today on LWN an article about ext4 gaining the ability to do
> > per file/folder case-insensitivity[1]. I can see some value in this
> > property existing for subvolume/folder/file level for Btrfs for things
> > like Wine and Darling, which could take advantage of this to help
> > support Windows and macOS applications that expect insensitivity to
> > work properly. In particular, I'm looking at how games are glitchy
> > with case sensitivity because it's rarely tested (both Windows and
> > macOS are case-insensitive by default).
> >
> > Has anyone looked at what it would take to do this in Btrfs too?
> >
> > [1]: https://lwn.net/Articles/829737/
>
> It might also make virtiofs on Btrfs more compelling across
> virtualized Windows and macOS.

i.e. virtiofs pointed at a subvolume; suggesting if possible making
the behavior enabled not just be feature flag (if required) but also
xattr set on the enclosing subvol.


-- 
Chris Murphy

Re: Optional case-insensitivity

[Mark Harmstone]

On 28/8/20 8:43 pm, Neal Gompa wrote:
> Hey,
>
> So I saw today on LWN an article about ext4 gaining the ability to do
> per file/folder case-insensitivity[1]. I can see some value in this
> property existing for subvolume/folder/file level for Btrfs for things
> like Wine and Darling, which could take advantage of this to help
> support Windows and macOS applications that expect insensitivity to
> work properly. In particular, I'm looking at how games are glitchy
> with case sensitivity because it's rarely tested (both Windows and
> macOS are case-insensitive by default).
>
> Has anyone looked at what it would take to do this in Btrfs too?
>
> [1]: https://lwn.net/Articles/829737/
Hi Neal,

FWIW, I emulate the reverse of this, optional case-insensitity, in my Windows
Btrfs driver, by using the xattr user.case_sensitive.

To do it properly, presumably all you'd need to do is add another DIR_INDEX/
DIR_ITEM pair to the inodes in question, though it'd imply another compat_ro
flag. I think it's something I've mentioned to Chris Mason before now.

Mark