Jennifer Pioch wrote:
> On 12/5/08, April Chin <April.Chin at sun.com> wrote:
> > Thank you for the code reviews...
> >
> > The nightly builds with the latest changes are still running, but in the
> > meantime, here are webrevs for the latest changes for ksh93 update1,
> > in response to code review comments so far.
> >
> > For code reviewers of the last changes, posted Nov 28th, here is a
> > diff against the last webrev:
> > http://cr.opensolaris.org/~chin/webrev-nonast.nov28-vs-dec4-diffs/
> >
> > A current webrev of all the non-AST changes to be reviewed,
> > diff-ed against the OS-Net gate, is at:
> > http://cr.opensolaris.org/~chin/webrev-nonast.dec4
>
> I'm vehemently opposed to the changes in
> usr/src/uts/intel/shbinexec/Makefile and
> usr/src/uts/sparc/shbinexec/Makefile: I think the shbinexec module
> should stay in /kernel and not /usr/kernel. So please revert this
> change
This doesn't make sense right now since there is no ksh93 in the root
filesystem by default in Solaris Nevada - the kernel module would work
if /usr/ is not mounted but there is no shell interpreter to execute.
That's why I moved it from /kernel/ to /usr/kernel/ for now. And there
is no point of "locking horns" since the compiled kernel module can be
_moved_ to /kernel/ without problems (note: I'm going to "lock horns"
with anyone who tries to challange the entry for "/sbin/ksh93" in
shbin.c) which you currently have to do anyway with { libshell, libcmd,
libdll, libsum and libast } for the non-Nevada OpenSolaris
distributions.
The whole ksh93-in-the-root-filesystem situation will only change if we
can get an approved ARC case which moves ksh93 to the root filesystem -
and that requires at least one consumer (e.g. project) who really wants
ksh93 in the root filesystem, lives in OS/Net and targets Solaris
Nevada.
----
Bye,
Roland
--
__ . . __
(o.\ \/ /.o) roland.mainz at nrubsig.org
\__\/\/__/ MPEG specialist, C&&JAVA&&Sun&&Unix programmer
/O /==\ O\ TEL +49 641 3992797
(;O/ \/ \O;)
