Re: globals vs parms
Matthias Andree <[email protected]>
| Newsgroups | gmane.mail.bogofilter.devel |
|---|---|
| Message-ID | <[email protected]> |
David Relson <[email protected]> writes: > I can live with those problems. Transaction support requires that > there be a unique database environment. For bogoutil, which only works > with one database, uniqueness is guaranteed. Uniqueness also doesn't > matter if transactions aren't being used. > > You don't like my lightweight solution for determining if the > uniqueness check is needed. If you've got a comparably simple way of > doing the check, let me know! My first concern is that the path of the decision is blurry. We export what program we're running, but what this means, isn't clear without checking every "consumer" of the new variables. If we are using globals, we should make their semantic clear. If there's a global variable to defeat - for instance - the check that we're using only one environment, because we know that the program guarantees this in another way, for instance, by ignoring --wordlist parameters, then the variable should do just that, and be called accordingly. fBogolexer says next to nothing. -- Matthias Andree _______________________________________________ Bogofilter-dev mailing list [email protected] http://www.bogofilter.org/mailman/listinfo/bogofilter-dev