[U-Boot] [RFC] Codepages for console, video, FAT file system

[Simon Glass]

Hi Heinrich,

On Thu, 30 May 2019 at 04:44, Marek Vasut <marex at denx.de> wrote:
>
> On 5/30/19 12:38 PM, Heinrich Schuchardt wrote:
> > The Linux console typically is running with Unicode. So with a serial
> > connection we are transferring UTF-8. In the UEFI sub-subsystem we also
> > assume that UTF-8 is used for the console.
> >
> > Linux typically mounts FAT file systems assuming codepage 437 for the
> > filenames. Windows assumes codepage 1250.
> >
> > Currently when accessing the FAT filesystem the console characters are
> > used verbatim without conversion to codepage 437 nor 1250 which leads to
> > incorrect filenames when the partition is mounted in Linux or Windows.
> >
> > The drivers/video/vidconsole-uclass.c has a comment that it is assuming
> > codepage 437.
> >
> > Some keyboard drivers also assume codepage 437.
> >
> > My suggestion to overcome the chaos is that we change all our coding to
> > assume that the console talks UTF-8 and the FAT filesystem gets a
> > customizing setting defining if it is mounted as codepage 437 or 1250.
>
> Sure, if it doesn't add bloat into SPL and/or grow the binary for
> systems not using UEFI.

I don't like the idea of using unicode everywhere. I worry about
confusion, complexity and massive code bloat.

What do we need unicode in the console?

Regards,
Simon