On Wed, Dec 23, 2009 at 11:15:53AM -0800, Garrett D'Amore wrote: > James Carlson wrote: > >Garrett D'Amore wrote: > >>Btw, I'm of the mind that it may be questionable to retain the old v1 > >>mapfile syntax for very long. As indicated, not many people are using > >>it, and we really shouldn't have to carry around baggage ~forever. I > >>don't think the mapfile syntax was ever officially part of our source > >>compatibility story. :-) > > > >I think that's a step too far. Yes, they're Committed interfaces and, > >no, it would not be good to have them go away. > > But wouldn't a tool to convert them answer the requirement for Committed > support? I don't think the specific compiler command lines are part of > that Committed interface, but perhaps I'm wrong. And I think I'd propose > marking them Obsolete (at least) as part of this.
Third parties' builds would still break until the mapfiles are converted. Yes, they could then do that. That's just painful though. I agree with Jim, removing v1 mapfile support is a step too far. I would prefer that the default somehow be that when building executables on OpenSolaris the stack ends up being not executable. (It'd also be nice if ld could warn about executable stacks when no mapfile explicitly set the stack to be executable. And it'd be nice if the warning said what to do about it. I'd settled for this if the default just could not be made that new executables get non-executable stacks.) Nico --
