On Fri, Feb 14, 2014 at 12:36:15PM -0500, Kevin A. McGrail wrote: > >>> This warning is more for the rule developers so that they can put > >>> "if (version >= 3.003002)" around the problematic rule. If > >>>there was some > >>> way to suppress it except during a full make, that might be a better > >>> solution than version checks. Would dbg() do that? > >> > >>Good point. I think you can remove the code. It won't apply in > >>anything after 3.3.2 clearly. > > > >Do we want to be aware of creating rules that may cause problems > >for pre-3.3.2 installs? As I said, the primary use of that warning > >is for rule devs. > > > >Or is sa-update on trunk no longer feeding rule updates to 3.3.1 > >and older versions so it truly is no longer an issue? > > > I will add this to my list o' rule gen questions as I believe we are > generating rules. And I'm sorry I read this poorly and now > understand the debug is so developers use if version loops. > > Do people really care if the debug output is long? >
From my (Debian's) point of view, sa-compile is run daily via cron if necessary following an sa-update run. sa-compile runs with --quiet, and I'm happy if these warnings are suppressed only under this condition. It'd be really confusing if these messages were generated on a regular basis on a production system. Rule developers testing their rules should see these in most cases, so it seems reasonable to show the warnings under normal operations. Would it also make sense to enclose the rules that trigger this warning in a version check, so we can avoid it in the standard rule set? I'll put together code to suppress these messages under '--quiet'. noah
signature.asc
Description: Digital signature
