This is the mail archive of the
newlib@sourceware.org
mailing list for the newlib project.
RE: [PATCH] stdlib/sb_charsets.c: Don't leave untranslatable characters
> That was apparently only half the truth. My assertion was based on
> testing three codepages. A more thorough check revealed that
> in certain
> codepages the undefined byte values are not converted to
> their identical
> UTF-16 values. Instead they are converted to unambiguous UTF
> values in
> the UTF private use area between U+f700 and U+f8ff. I don't
> know why that
> is, but I checked that the values are identical over different Windows
> versions (Windows 2000, Windows XP, Windows 7), as far as the
> codepages
> are available on the machines.
This rings a bell: I have been told that the SFU/Interix product stores filename characters unsupported by NTFS (the Windows Explorer tells me that they are \, :, *, ?, >, <, |, ", and /, which doesn't count) as private use characters.