On Sun, 2009-06-07 at 11:10 +0200, Vladimir 'phcoder' Serbinenko wrote: > On Sun, Jun 7, 2009 at 4:18 AM, Pavel Roskin<pro...@gnu.org> wrote: > >> Improving > >> strcasecmp is possible and may even be compact. Even if unicode counts > >> a lot of alphabets only few are bicameral. AFAIK main ones are Latin, > >> Greek, Cyrillic and Armenian. I hope that in most cases the lowercase > >> conversion can be done with some simple arithmetic operations > > > > We are using grub_strcasecmp() on data in other encodings in some cases. > > I think short names on the FAT filesystem are never in UTF-8. > Normally not but fat code passes short filenames to grub without any > encoding conversion. In other words fat code assumes short filenames > are in utf8. A problem in both hfs and fat is that it uses local > encoding and at least in case of fat it's impossible to know which one > from filesystem alone. We would need something like mount option to > get this right. It could be sth like > hd0_0_codepage=437
Fortunately, it would only matter if the user tries to access a file with non-ASCII characters using a capitalization different from the one used by the filesystem. > > Also, > > case comparison depends on the locale. That's too much complexity for > > the code GRUB. I would prefer the case comparison for hfsplus to be in > > the hfsplus module if we decide to implement it correctly. > I agree with you. I see no reason for users to have unicode characters > in kernel names. And so getting it right is too much work for an > almost zero benefit I case of hfsplus, there is a chance to miss a file in the binary tree if there are non-ASCII files in the same directory. But I agree, it's not worth the trouble at this point. -- Regards, Pavel Roskin _______________________________________________ Grub-devel mailing list Grub-devel@gnu.org http://lists.gnu.org/mailman/listinfo/grub-devel
