Garrett D'Amore wrote: > Peter Memishian wrote: > > > i'm afraid i might be damming myself by making this suggestion, > > > but would using a pre-compiled ksh script solve this problem? > > > > I'm still puzzled by the whole need to replace `sleep' with a built-in. > > What was wrong with `sleep' as a C program? The purpose in life of sleep > > is to do nothing for a specified period of time; why does the performance > > of it matter? > > > > >From where I sit, it seems this transition has already cost us too much. > > Maybe it's time to re-evaluate. > > > > > I think one of the points was, surprisingly, that there were additional > features in ksh93's sleep. (Fractional sleeps, e.g. sleep for 500 msec > instead of 1 sec granularity.) > > If ksh93 already has the code, and *if* we accept that it is now > entrenched into ON (such that one can't have a properly functioning ON > without it) as a critical piece, then reducing duplicated code fragments > *elsewhere* might be worthwhile in itself. > > (I've already complained about a number of things -- such as increased > start up times of shell utilities -- I don't like about the ksh93 but > was either ignored or overruled.
Erm... you weren't ignored nor overruled. The discussion lead to the development of the "shbinexec" kernel module which recognizes precompiled shell code and runs the matching interpreter (and bypasses the whole isaexec machinery (which is AFAIK your main concern)). The final result is that the matching tools start and run faster. > Seems a bit late to restart the whole > ksh93 debate at this juncture...) See above... ---- Bye, Roland -- __ . . __ (o.\ \/ /.o) [email protected] \__\/\/__/ MPEG specialist, C&&JAVA&&Sun&&Unix programmer /O /==\ O\ TEL +49 641 3992797 (;O/ \/ \O;) _______________________________________________ opensolaris-code mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/opensolaris-code
