Re: CVS commit: src/sbin/gpt

[John Nemeth <jnemeth%cue.bc.ca@localhost>]

On Nov 29,  4:16pm, Christos Zoulas wrote:
} In article <201511291418.tATEIwRt002398%server.CornerstoneService.ca@localhost>,
} John Nemeth  <jnemeth%cue.bc.ca@localhost> wrote:
} >
} >     Why did you leave out the call to getdisksize() for when the
} >ioctls don't exist?  Actually, why did you even bother with this
} >patch, since it is pretty much equivalent to what was there before,
} >minus the missing functionality?
} 
} Because who has this functionality (drvctl) and does not have the
} ioctls? Only the NetBSD-6 tools build. This is too much complexity
} for very little benefit. It is simpler to add the ioctls to NetBSD-6
} than keeping all this messy libprop code around (which we have again
} in fsck...) :-)

     The reason I noticed it was due to a build failure.  The system
I was building it on is:

NetBSD msgate 7.0_BETA NetBSD 7.0_BETA (GENERIC.201408231540Z) amd64

(Yes, it needs to be updated, but that's beside the point.)  Granted,
this did point out an additional problem, which is:  why was it
using the system headers and not the headers from the source tree
(a full cvs update was done immediately before the build)?

     BTW, if you're going to take out getdisksize(), then there is
really no point in having #ifdef DIOCGSECTORSIZE, since it won't
work even if it does compile.

}-- End of excerpt from Christos Zoulas