Compilation problem with bogofilter-0.10.1.3 - tests still fail.

[David Relson]

Nick,

A preliminary result:  t.robx reported finding 1144 words in 
tests/bogofilter/inputs/spam.mbx which is the spam training database and 
3073 words in good.mbx.  The expected countes are 1144 and 3074.  So, the 
lexer on your machine missed a word.

As a second fact, "t.robx -v" also writes file output.vvv which includes 
all tokens in the spam.mbx.  It's conspicuously different from mine because 
mine is alphabetized and yours is not.  This indicates a (probably 
irrelevant) difference in our versions of BerkeleyDB.  More in a 
while.  I've got to go run some errands.

David