> Concerning fork/vfork/posix_spawn, I have observed > good performance enhancements by switching from fork > to vfork and again from vfork to posix_spawn in > "xjobs".
Strange that you had that much difference between vfork and posix_spawn, I thought the second one was just a wrapper to make using vfork safe. > So I would really vote for a change of > implementation for creating child processes within > shells, especially as I suspect that this could > enhance performance of "./configure" runs etc. That is what I had in mind. > Have you contacted the maintainer of tcsh regarding a > switch from fork/vfork to posix_spawn? So maybe if > tcsh ever gets updated within Solaris, this > enhancement will be included indirectly... See this thread: http://mx.gw.com/pipermail/tcsh-bugs/2006-November/000455.html apparently they prefer to keep fork as the default for safety, but believe the vfork version should work just fine. So it seems solaris should either stick to fork, or switch to vfork long enough before a release to get some appropriate testing done. This message posted from opensolaris.org _______________________________________________ opensolaris-code mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/opensolaris-code
