Mike Kupfer wrote: > > >>>>> "JP" == Jennifer Pioch <piochjennifer at googlemail.com> writes: > > JP> I'm vehemently opposed to the changes in > JP> usr/src/uts/intel/shbinexec/Makefile and > JP> usr/src/uts/sparc/shbinexec/Makefile: I think the shbinexec module > JP> should stay in /kernel and not /usr/kernel. > > I don't feel strongly about the location of the shbinexec module. I do > feel strongly that if the module lives in /kernel, someone from the > ksh93 team needs to verify[1] that there aren't any issues related to > the boot archive.
I did a quick look and the kernel module is 4656 bytes on 64bit SPARC, AFAIK this is not going to cause big trouble for the boot archive. But as I said in http://mail.opensolaris.org/pipermail/ksh93-integration-discuss/2008-December/006634.html it doesn't make sense to have it in /kernel/ without moving ksh93 and the libraries, too - and that needs an ARC case with a "bulletproof consumer" (e.g. a project which wishes to use ksh93, can't easily work around the lack of ksh93 and is willing to endure the pain of such an ARC case (that's basically the outcome of PSARC/2006/550 where /sbin/ksh93 was removed because we didn't had a consumer)). For now we have the option that OpenSolaris distributions can manually move the compiled module from /usr/kernel/exec/sparcv9/shbinexec to /kernel/exec/sparcv9/shbinexec, put a copy of the ksh93 frontend into /sbin/ksh93 and move the ksh93 libraries from /usr/lib to /lib and it will work on the root filesystem (that's why "shbin.c" as the entry for "/sbin/ksh93" as the last option in the list). ---- Bye, Roland -- __ . . __ (o.\ \/ /.o) roland.mainz at nrubsig.org \__\/\/__/ MPEG specialist, C&&JAVA&&Sun&&Unix programmer /O /==\ O\ TEL +49 641 3992797 (;O/ \/ \O;)
