On Mon, Dec 20, 2004 at 03:59:42AM -0800, [EMAIL PROTECTED] wrote: > This seems to go wrong with some character encodings. I think this is > probably because I receive the data as a byte array and the conversion > to a string is not intelligent enough. For example if the input data is > UTF-8, I suspect I am converting each byte of a multi-byte character > sequence into a UTF-16 character.
Yes, nsIScriptableInputStream sucks like that... Note that scriptableinputstream gives you what it calls a string. Which is a problem if it contains a null byte (e.g. UTF-16), because the string will be truncated there... > So is there a scriptable way of converting a byte array in a known > encoding to a correctly-decoded Javascript string? Or is there another > way I can register as a listener to get Unicode data? http://lxr.mozilla.org/seamonkey/source/intl/uconv/idl/nsIScriptableUConv.idl Note that the byte array methods got added after Mozilla 1.7.x; they are only in 1.8a5. Since nsIScriptableInputStream.read gives you a string though, that should be ok(-ish) unless you use UTF-16, UCS-2 or UTF-32. If you want to read a byte array instead of a string, you can use nsIBinaryInputStream.readByteArray. -biesi -- _______________________________________________ Mozilla-netlib mailing list [EMAIL PROTECTED] http://mail.mozilla.org/listinfo/mozilla-netlib
