2011/2/4 Karsten Bräckelmann <[email protected]>: > On Fri, 2011-02-04 at 04:15 +0100, Karsten Bräckelmann wrote: >> On Thu, 2011-02-03 at 16:35 -1000, Warren Togami Jr. wrote: >> > I am not sure who "smf" is, but are they aware that "test rules" without >> > "tflags nopublish" could possibly be auto-promoted with a score they >> > don't expect. Scores in the sandbox below are ignored. We don't want a >> > repeat of the FSL_RU_URL issue. >> >> Getting *cough* late here, and I'm terribly tired... > > Argh. That's what you get for searching at the wrong end. Right after > turning off the machine, and heading to bed... > > That's Steve Freegard, a recent addition to the committers. :) > > Anyway, I stand to my recommendation to just add the tflags. This is > something we all need to get into the habit > >> Warren, IMHO, please feel free to add the tflags necessary. Yes, inside >> someone else's sandbox. I guess the recent discussion about rules >> possibly spreading unintended didn't reach everyone. Although it won't >> right now, as we know -- but the corpora being under-limit is another >> topic, unrelated to best practices. Adding tflags nopublish is not >> intrusive anyway, and warranted in this case, I guess.
+1 btw, iirc it's possible to set "tflags nopublish" on a file-by-file basis -- add "#testrules" at the top of the file, and all rules after that point are implicitly nopublish. see rulesrc/sandbox/jm/20_bug_5984.cf for an example. https://issues.apache.org/SpamAssassin/show_bug.cgi?id=5545 has the background. --j.
