> To build web stack , please refer to the > documentation here > http://hub.opensolaris.org/bin/view/Project+webstack/s > unwebstack
Those docs seem *much* more complete than whatever I had before (I've misplaced the link). Thanks for pointing them out. The only other build reference I found was the SFW link from http://hub.opensolaris.org/bin/view/Project+webstack/files . >I can definitely > assure you that performance gain around 2-5% will > definitely come under > noise if you add database to this equation. I believe you're right. Although, I HAVE noticed one oddity. I have a simple "benchmarking" script the developers have handed me. It's really just a script that does nothing but: for($i = 1; $i <= 2000000; $i++) { $x=$i * 5; $x=$x + $x; $x=$x/10; $string3 = $string1 . strrev($string1); $string2 = substr($string1, 9, 1) . substr($string1, 0, 9); $string1 = $string2; } (I know this is a bad benchmark, but it's what the devs handed me...) If I compile plain PHP from source - on my OpenSolaris x86 machine, using SS12u1 - that script has an in-code (not startup) average measured time of 305 ms. If I recompile, adding nothing but "--with-pdo-pgsql=shared,/usr/postgres/8.2" to the configure flags, the average time jumps to a consistent 379 ms. Keep in mind that this is the script measuring itself, using microtime (), and I'm running php right where it was built... not installed at all. How odd! This definitely isn't a Webstack thing, nor a Solaris/OpenSolaris thing, and I'm glad it's not a SPARC platform thing, since I can reproduce it on my Intel workstation. In fact, I have no idea who to even ask about this. -- This message posted from opensolaris.org
