mbtowc - combining character

[Old Wolf]

As far as I can see, mbtowc and mbstowcs assume that there is
exactly one wide character for each multi-byte sequence. How are
you meant to cope with MBS that correspond to two wide characters?

For example, if it is Unicode and the MBS represents a letter
with a combining diacritic.

 

[klaushuotari]

As far as I can see, mbtowc and mbstowcs assume that there is
exactly one wide character for each multi-byte sequence. How are
you meant to cope with MBS that correspond to two wide characters?

For example, if it is Unicode and the MBS represents a letter
with a combining diacritic.

You aren't. That's purely implementation defined.

 

[CBFalconer]

As far as I can see, mbtowc and mbstowcs assume that there is
exactly one wide character for each multi-byte sequence. How are
you meant to cope with MBS that correspond to two wide characters?

For example, if it is Unicode and the MBS represents a letter
with a combining diacritic.

The same way you convert '\n' to a cr/lf output sequence.

 

[Guest]

As far as I can see, mbtowc and mbstowcs assume that there is
exactly one wide character for each multi-byte sequence.

There is exactly one wide character for each multi-byte sequence.

How are
you meant to cope with MBS that correspond to two wide characters?

For example, if it is Unicode and the MBS represents a letter
with a combining diacritic.

Those are two separate multi-byte sequences. The C functions work on
the character level, not on the glyph level.

 

[Boudewijn Dijkstra]

Op Wed, 04 Apr 2007 01:18:46 +0200 schreef Old Wolf

As far as I can see, mbtowc and mbstowcs assume that there is
exactly one wide character for each multi-byte sequence. How are
you meant to cope with MBS that correspond to two wide characters?

For example, if it is Unicode and the MBS represents a letter
with a combining diacritic.

Perform canonical decomposition before converting.