Danek Duvall wrote: > > James C. McPherson wrote: > > > http://cr.opensolaris.org/~jmcp/6414832/webrev > > Is the new way of computing POUND_SIGN in Makefile.master on top of setting > POUND_SIGN in nightly and bldenv just belt and suspenders? Shouldn't the > first do the trick? (Cute trick, BTW.) > > gen_make.machines.sh: > > - line 30: I'd just set PATH=/usr/bin:/usr/sbin and not bother with > setting variables for all the commands, especially since you're not > allowing folks to override them from the outside. > > - line 39: Since you're in ksh93, you can skip running expr and use the > more readable math expression: > > max=$(((cpus + 1) * 2))
Please change that to (... and make sure you correctly declared the variables with either "integer max cpus" or "float max cpus" (depending on what you need (BTW: ksh93 supports ISO C99 math functions for rounding, too))) ... -- snip -- (( max=((cpus + 1) * 2) )) -- snip -- Using var=$((expr)) means that the internal representation of "expr" is converted to a string (and if "var" is declared as integer or float you get a string-to-integer/float conversion, too) while the form above keeps the internal value intact (the string-to-integer and integer-to-string conversions eat performance and cause rounding errors if floating-point values are used (because strings use base10 while the interbal IEEE754-2008 representation of a |long double| uses base2)). ---- Bye, Roland -- __ . . __ (o.\ \/ /.o) roland.ma...@nrubsig.org \__\/\/__/ MPEG specialist, C&&JAVA&&Sun&&Unix programmer /O /==\ O\ TEL +49 641 3992797 (;O/ \/ \O;) _______________________________________________ tools-discuss mailing list tools-discuss@opensolaris.org
