>>>>> "Paul" == Paul Rohr <[EMAIL PROTECTED]> writes:
Paul> At 09:39 PM 2/3/01 GMT, Jesper Skov wrote:
>> This fixes bug 223 (and possibly 436 - haven't tested yet). Problem
>> was that fields (objects in general) got encoded as spaces to the
>> buffer used for finding word limits. I replaced the spaces with
>> the letter A (using a macro) and added references to the bug + some
>> explanations.
Paul> Jesper,
Paul> I'm no Unicode expert, but given the intended semantics of
Paul> UCS_OBJECT, you may be using the wrong codepoint:
Wow. I just named it like that - it's not the proper unicode value I
use. I considered using UCS_ABI_OBJECT but decided against it since
some of the other tags seemed to be AbiWord specific as well. But I
guess I could be wrong.
Paul> I'm not sure that it's much better than spaces in the current
Paul> context, though. The field width stuff mentioned later in this
Paul> thread may be a more promising alternative.
Oh, it's much better than spaces for the simple reason that spaces are
not word characters. I have hardwired the BELL to be considered a word
character.
It's arguably ugly as sin and I'd be happy to replace it with some
better suited unicode - but I didn't have the patience to go through
all the unicodes to find a suitable code. Hence the (should have been)
ABI specific macro so we could easily replace it with something better
suited when and if it was found...
Jesper
