On Wednesday 23 of February 2011 18:09:42 Russ Cox wrote: > > I'm unsure if this conversation is about Plan 9 or plan9port, but in > > any case I've used Local for lots of other things on Plan 9, > > particularly name space manipulations. There, I don't understand why > > it needs restrictions. > > > > Or are you just saying that on plan9port you need to do magic so you > > might as well catalog the tricks? In that case, I understand. > > Yes, the idea is that on plan9port you might change > the implementation of Local to pick off var=value > and cd path and run those internally. It could still > shell out for other commands like Local echo $foo.
How about reading /proc/$pid/environ (where $pid is the shell spawned for command execution) before the $pid exits and transfering all the environment variables back to the Acme's own environment? I'm not sure, but I think that /proc/$pid/environ should be accessible for Acme after exit() by the $pid, but before Acme's waitpid($pid, ...) completes. -- dexen deVries [[[↓][→]]] > how does a C compiler get to be that big? what is all that code doing? iterators, string objects, and a full set of C macros that ensure boundary conditions and improve interfaces. ron minnich, in response to Charles Forsyth http://9fans.net/archive/2011/02/90