On Jul 7, 2006, at 7:55 AM, Jonathan del Strother wrote: > On 7 Jul 2006, at 15:27, Jason A. Hoffman wrote: >> On Jul 7, 2006, at 5:16 AM, Jonathan del Strother wrote: >> Why does mongrel_rails insist on loading ~/.irbrc with each request? >>> a) I'm curious why it loads it at all (I assume there's no way of >>> getting an inline breakpointer??) >>> b) Why re-load it? It causes problems with any constants that are >>> used in .irbrc... (alternatively, how do I avoid re-assigning to a >>> constant?) >>> >>> Jon_______________________________________________ >> >> >> Are you starting it in development mode? > > Yep. And no, it doesn't load irb in production mode. I'm just > curious why it loads it at all. It does seem to be Mongrel (rather > than Rails) that's loading it - it doesn't get loaded when I run my > server off of lighty, for instance.
When you drill through a mongrel, webrick or a FCGI when in development mode you'll see stat64 calls and then that they're coming from irb (you don't usually see it with lighttpd because it hides all that from you). Like the table from http://blogs.sun.com/roller/page/bmc? entry=dtrace_on_rails Top ten system calls per URI serviced: -----------------------------------------------------------------------+ ------ setcontext | 15 fcntl | 16 fstat64 | 16 open64 | 21 close | 25 llseek | 27 lwp_sigmask | 30 read | 62 pollsys | 80 stat64 | 340 - Jason _______________________________________________ Mongrel-users mailing list [email protected] http://rubyforge.org/mailman/listinfo/mongrel-users
