[Bug localedata/13061] New: iconv mapping of 0xA8 0xEC in CP1258 isnon-canonical

["bruno at clisp dot org" <sourceware-bugzilla at sourceware dot org>]

http://sourceware.org/bugzilla/show_bug.cgi?id=13061

           Summary: iconv mapping of 0xA8 0xEC in CP1258 is non-canonical
           Product: glibc
           Version: 2.14
            Status: NEW
          Severity: normal
          Priority: P2
         Component: localedata
        AssignedTo: libc-locales@sources.redhat.com
        ReportedBy: bruno@clisp.org


Bug 12777 <http://sourceware.org/bugzilla/show_bug.cgi?id=12777>
was fixed to map U+0385 (like U+1FEE) to 0xA8 0xEC. Good.

But at the same time, in the reverse direction, 0xA8 0xEC ought to map to
U+0385, not to U+1FEE. Why?

1) http://www.unicode.org/charts/PDF/U1F00.pdf states
that the decomposition of U+1FEE is U+0385. That is, U+0385 is a "simpler"
Unicode character than U+1FEE, although both look very similar
(cf. http://www.unicode.org/charts/PDF/U1F00.pdf and
http://www.unicode.org/charts/PDF/U0370.pdf).

2) According to http://www.unicode.org/versions/Unicode6.0.0/ch07.pdf,
the block U+0370..U+03FF is more for modern Greek, whereas the block
U+1F00..U+1FFF is mostly for ancient Greek. But CP1258 is about modern Greek.

To reproduce:

$ printf '\xA8\xEC' | iconv -f CP1258 -t UCS-4LE | od -t x4
0000000 00001fee
0000004

Should be:

$ printf '\xA8\xEC' | iconv -f CP1258 -t UCS-4LE | od -t x4
0000000 00000385
0000004

Attached is probable fix (untested).

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.