1more work on test suite
2
3Real bugs:
41. unicode_iconv needs to save the shift state when backing up
52. the init function for a converter has no idea if it is initializing
6   for read or write.  should have two init functions
7
8Treat malformed (and overlong, except maybe c0 80) UTF-8 characters
9like 0xfffd.  Define for 0xfffd in converter header.
10
11we need some more functions:
12  a way to map character number to byte number
13    (so we can continue to use strncmp)
14  a strchr and strrchr using a unicode_char_t argument
15  strcasecmp and strncasecmp
16
17intermediate buffer:
18  * how to size it?
19  * should be circular for efficiency
20  * should be bypassed when converting to/from UCS-4
21
22more charsets
23  Japanese charsets like JIS are very important
24  These must be converted from Unicode tables using a script
25
26name charsets appropriately
27names of existing charsets should be changed.
28eg ucs2 -> utf16
29look at IANA registry
30
31look at Solaris and glibc 2.1:
32  make sure we use the same names for the same encodings
33  see if we have to make it so our aliases apply even when
34    using the system iconv (necessary when we want to use a fixed
35    charset but existing systems don't agree on what it should be
36    named)
37
38if fromcode==tocode, then just copy through
39
40merge with FriBidi
41
42use gtk-doc
43
44ucs readers which understand the feff convention
45
46turn converters into shared objects of their own and dynamically load
47them as requested
48
49unicode_canonical_decomposition has a bad interface
50Should let user give a buffer
51
52
53================================================================
54
55more properties:
56
57* mirrored.  use 2-level table
58