On Fri, Apr 28, 2006 at 06:31:55PM -0400, Laszlo (Laca) Peter wrote: > We shouldn't invent any more g-prefixed commands.
Agreed; the approach of introducing them only when they're in reasonably widespread use seems fine. > Some more questions: > > 4. What should be the update model for /usr/gnu? > > I think it will only be useful if it's kept up-to-date. Now we're teetering on the edge of the /usr/sfw abyss again. '/usr/gnu' doesn't have an update model. It does have a stability level, probably Stable. But that doesn't seem to be what you're asking. Each interface exported by components provided therein has its own stability level. Perhaps it makes sense as a general rule for GNU interfaces to be classified Volatile so that they can be updated arbitrarily. But there are certainly instances in which that's unnecessary and a higher stability level could and should be used (the canonical example is gzip, which is irrelevant since it's already in /usr/bin, but there are others as well). Enough about what types of updates are permitted; what about the updates that will actually be integrated? That depends on resources, and historically few resources have been allocated to this type of work. It sounds like you have an active interest in this, so perhaps it would make sense for you to perform the necessary updates. > 5. What should be the policy for using GNU commands in other > parts of Solaris? This also assumes that every interface exported by anything in /usr/gnu has the same stability level. Shouldn't we just say that the usual rules apply and leave it at that? -- Keith M Wesolowski "Sir, we're surrounded!" Solaris Kernel Team "Excellent; we can attack in any direction!" _______________________________________________ tools-discuss mailing list tools-discuss@opensolaris.org
