[PATCH 0/2] Native case-insensitive file-system support (try 2)

[Charles Davis]

Alexandre Julliard wrote:
> You certainly don't want to enumerate mount points on every file
> lookup.
The problem is that getattrlist() will only return volume attributes if
the file-system object you pass to it happens to be a volume. There's
probably a way to optimize it, but I've left it as-is for now. If you
prefer, I can remove the getattrlist()-related parts.

> Also the non case-preserving handling looks quite suspicious
> (and IMO useless),
I'll consider removing it. I've left it in for now.

The thinking behind it was that file systems that don't preserve case
just don't care what case the filenames have. If you ask for a file and it's
not there, it's not there.

> and you have to fix the FIXMEs.
Done.

> Plus there's no reason to make this Mac OS specific.
Fixed. The non-Mac OS functionality hasn't been tested yet, though,
because I don't have any Linux boxes or VMs.

Scott Ritchie wrote:
> Wine still needs to be made aware when it's on CIOPFS though (similar to
> how it behaves on fat32 partitions)
This patchset does that now.

> Charles Davis wrote:
>> It's easy enough to tell if you're on a
>> FUSE FS (statfs(2) will tell you that), but it's somewhat tedious to
>> tell exactly which FUSE FS it is (AFAIK, the only way to tell is to
>> read the mount table).
> Well we could do something blunt like have CIOPFS deliberately make a
> "hey I'm case insensitive file" that Wine could look for.
That's the approach I took in this patchset. I need to hook this up on
the other end, though (i.e. have the FS driver generate a fake file
named '.ciopfs').

Chip