Bogofilter migration & tuneup

[Matthias Andree]

Robin Bowes <robin-lists at robinbowes.com> writes:

> Robin Bowes said the following on 05/12/2005 14:31:
>> I'm now trying bogoutil -d wordlist.db > wordlist.txt again to see if it
>> grows massively again.
>
> wordlist.db is 2.1GB and growing (from a 105MB wordlist.db).
>
> Does that sound like it's corrupted?

Yes. .db and .txt files are roughly the same order of magnitude in
size. A factor of 10 looks corrupt. You can use db_verify to see what it
thinks of your wordlist.db (don't run any registration at the same time,
else you'll see false positives). As db-recover hasn't helped you, and
db-recover-harder is no longer workable since the log files have been
removed, you'll probably have to use db_dump with "-r" (or perhaps "-R"
if necessary, but that may print junk) option to see what you'll get and
if it's reasonably complete for db_load and a subsequent "bogoutil -d"
operation.

The detailed salvaging instructions were not shipped with 0.93.5 but
joined in with some 0.94 versions, please see sections 3.2 with all
subsections of the newer README.db:

<http://cvs.sourceforge.net/viewcvs.py/bogofilter/bogofilter/doc/README.db?rev=1.36>

In section 3.2.4, use the new bogofilter 1.0.0 version for "bogoutil
[whatever] -l".

-- 
Matthias Andree