Steffen Nurpmeso dixit: >and when running it it seems that mksh(1) (this is @(#)MIRBSD KSH >R51 2015/07/10 packaged for ArchLinux) uses a (i'm guessing its >edit) line buffer until that overflows the first time, before >passing the final data through to STDOUT:
Excuse me, I do not understand that paragraph. > ?0[sdaoden@wales tmp]$ time bash t.sh >/dev/null > 0m0.05s real 0m0.03s user 0m0.01s system > ?0[sdaoden@wales tmp]$ time dash t.sh >/dev/null > 0m0.02s real 0m0.01s user 0m0.01s system > ?0[sdaoden@wales tmp]$ time mksh t.sh >/dev/null > 0m6.76s real 0m2.04s user 0m4.28s system There seems to be no data not sent to stdout? Confused, //mirabilos -- <igli> exceptions: a truly awful implementation of quite a nice idea. <igli> just about the worst way you could do something like that, afaic. <igli> it's like anti-design. <mirabilos> that too… may I quote you on that? <igli> sure, tho i doubt anyone will listen ;)
