On 11.09.17 16:34, Rob Clark wrote:
On Mon, Sep 11, 2017 at 10:13 AM, Alexander Graf agraf@suse.de wrote:
On 09.09.17 12:47, Rob Clark wrote:
We'll eventually want these in a few places in efi_loader, and also vsprintf.
Signed-off-by: Rob Clark robdclark@gmail.com
UEFI wants UCS-2 (the 16 bit only Windows thing), not UTF16 (the may-also-be-32bit-wide thing).
So I'm not sure how far away this code is from the ultimate truth of what we need for UEFI compatibility. Do you have any gut feeling?
- leif & peter
I'm not entirely sure. Grub seems to treat it as UTF-16, and from a 30sec look at wikipedia, it seems like UCS-2 is a subset of UTF-16. But I won't claim to be a unicode expert.
I agree, but we should make sure to have our naming straight. If we only support the BMP (basic multiplingual plane) part of UTF-16, we should
a) Indicate that in the function names b) Bark if we get input that's outside the range?
Otherwise sooner or later people will get confused, as UTF-16 strings can for example be shorter than the sum of 16-bit-fields they occupy.
Alex