Re: Makeing the subshell reliable

[Oswald Buddenhagen <ossi kde org>]

On Thu, Jul 20, 2006 at 12:00:58PM +0300, Pavel Tsekov wrote:
> I hope this explanation helps. If you have questions, please, ask.
>
as i understand it, it is a pure implementation problem and not
something fundamental.
mc gets the shell's sigchld at any time (just as it does with output),
so it should be able to revive the shell at any time as well (a
background exit with a potential cwd update should be communicated
somehow - no, not with a popup, but maybe changing the prompt's
background color until one presses ctrl-o or something).
regarding the sigstop before/after pwd question ... i think it doesn't
actually matter. sigchld from the shell means it's ready. then we should
- purge it's input if we were in panel mode, so it does not start with
  the next command under us. we do this. you are suggesting that this
  can't work - why?
- get to know it's cwd. given that the cwd should fit in a pipe's buffer
  (on linux, maxpath is 4k, just as the pipe buffer - i don't expect that
  relation to violate my assumption on any other system), there is no
  need to select() on the pipe until the sigchld arrives. so whether we
  revive the shell first and read the dir afterwards or the other way
  round should not really matter. regarding /proc, i don't see an actual
  advantage in using it ...


-- 
Hi! I'm a .signature virus! Copy me into your ~/.signature, please!
--
Chaos, panic, and disorder - my work here is done.