Mike Kupfer wrote: >>>>>> "Rich" == Richard Lowe <richlowe at richlowe.net> writes: > >>>> o pkgdefs/common_files/i.ttysrch >>>> o pkgdefs/common_files/i.locallogin >>>> o pkgdefs/common_files/i.localprofile >>>> o pkgdefs/common_files/i.sdconf_i386 >>>> o pkgdefs/common_files/i.devlinktab_i386 >>>> o pkgdefs/SUNWrmodr/i.master >>>> o pkgdefs/common_files/i.confupdate >>>> o pkgdefs/common_files/i.etcprofile >>>> >>>> Class action scripts that each know how to bump the (sccs-style) >>>> ident string of an installed file on merging. > > Rich> I'm still not certain whether these would be better landing in the > Rich> gate (and thus filed in bugster) or the project gate (and thus > Rich> grommit). Thoughts? > > So these are all pasting the SCCS (and sometimes copyright) information > from the latest distributed version of the file into the installed > version of the file. Since the files will continue to have SCCS IDs as > long as the gate is in Teamware, I think any changes here should deliver > into the project gate.
That makes sense. > The one question in my mind is whether we want to proactively remove the > #ident line from the source file at the same time we remove the update > logic from the class action script. I'm not bothered either way. On the one hand, they are intimately connected to the class action script - so it seems reasonable to take them out now. But on the other hand, once we take away the update logic from the class action script - they will just be source file #ident lines, not unlike every other source file in the gate. (i.e.: I don't think they will interfere or get in the way of anything if we leave them there) So I'm fine with leaving them... cheers, steve -- stephen lau // stevel at sun.com | 650.786.0845 | http://whacked.net opensolaris // solaris kernel development
