Tomas,
Thanks for forwarding this. As we all know, I've long been hoping someone
could get some hard data on which hash flavors are found in the wild:
http://www.abisource.com/mailinglists/abiword-dev/01/February/1477.html
Actually, if I read the sources correctly, we're currently expecting 8/N/100
dictionaries (for N <= 64), which is why Henrik was able to get us working
with many of the locales mentioned at the bottom of Ramon's note --
particularly the ones mentioned here:
abi/test/wp/Spelling.abw
The problems Vlad diagnosed were for locales that had 7-bit dictionaries:
http://www.abisource.com/mailinglists/abiword-dev/01/March/0769.html
As mentioned in my note to Thomas, I continue to believe that a
slightly-smarter hashtable loader could do enough additional work at runtime
to detect which hashtable variant was stored on disk and map them all to an
appropriate variant in memory:
http://www.abisource.com/mailinglists/abiword-dev/01/April/1030.html
However, we won't know until it's coded. I personally am inclined to let
Thomas chew on this some more and see what he comes up with. :-)
Paul
