As per my previous post, FMP 8.5 provides for UTF 8 export.
On Nov 27, 2006, at 4:14 PM, Rowland Carson wrote:
At 2006-11-27 10:45 +0100 Beatrix Willius wrote:
I would say that UTF16 is your best bet. Does mySQL expect the
byte order mark (BOM)?
Trixi - thanks for your comments. I can't immediately find any
reference to what MySQL expects, but no doubt it's documented
somewhere.
Does FMP write it?
A TextWrangler hex dump of a sample file tab-text export as UTF-16
shows the first 2 bytes as FE FF followed by the data in 2-byte pairs:
0000: FE FF 00 36 00 39 00 36 00 09 00 31 00 30 00 39 ...6.9.6...1.0.9
Opened as text this shows 696 <HT> 109
This is a sequence of three bytes specifing the kind of UTF. You
can check with a hex editor, if it's there or not. Another idea:
UTF16 can be little or big endianness. Perhaps FMP and mySQL have
different expectations in this regard.
It appears to be 2 bytes in this case. I note that TextWrangler
includes several such flavours in its options for character
encoding. It identifies the FMP export file as UTF-16. The other
UTF-16 choices are UTF-16 No BOM, UTF-16 Little-Endian, UTF-16
Little-Endian No BOM.
Just to check, I created a similar file in TextWrangler as UTF-16
and hex dumped it to compare with the FMP export; the first
sequence of bytes was identical.
I wonder why FMP does not provide for UTF-8 in text exports,
although it does cater for it in imports.
regards
Rowland
--
| Wilma & Rowland Carson http://home.clara.net/rowil/
| <[EMAIL PROTECTED]> ... that's Rowland with a 'w' ...
