Hello, I'm in the TextConverter mud trying to test my
ByteTextConverter refactoring.
I've corrected it, so don't use the version in the inbox.
I still have one test failing due to a HACK in MacRomanTextConverter>>#toSqueak:
The hack converts 165 (16rA5) supposed to be a Bullet (U+2022) to
charCode 183 (16rB7) (centered dot I presume)
see also
http://unicode.org/Public/MAPPINGS/VENDORS/APPLE/ROMAN.TXTThen I just realized that pharo did implement the same
ByteTextConverter idea but with a strict unicode mapping for
MacRoman... (No CP1252 hack)
I guess this comes from Sophie and that they got it right, but this
has more implication with fonts leadingChar etc...
Anyway we used same reference...
the hack is not applied in the other direction (fromSqueak:) GRRR that
makes my test fail...
SO I SUPPRESS IT ! NO REGRET ?
Other remarks:
I don't really know what to do with invalid characters...
- Error
- answer, 0, -1, nil, 16rFFFD...
Euro character is encoded strangely in squeak (it does not use the
iso8859-1 charCode, but replaced currency glyph ยค as in MacRoman
encoding...)
and Character euro still answer the MacRomanEncoded euro (219 16rDB)
it should be the currency glyph or the true euro unicode...
Nicolas
Locked
