Re: Windows env-Bug?

On Jul 28, 11:28 am, Gerhard Reithofer <gerhard.reitho...@tech-
edv.co.at> wrote:

Controlling (and not "monitoring") the process environment is one of the
basic methods of the "interprocess communication", its one way to
forward informations to subprocesses and change information on current
process level.

It's the most basic method, yes: unidirectional (from parent to
child), and one-shot (only at spawn time).

But this part is not threatened by Joe's proposal: on the contrary, he
suggests to keep only this part (fill ::env with tclsh's environment
at interp init time, and update the environment of children based
on ::env's contents, only at spawn time; nothing in between. Hence no
cpu-hungry traces on ::env).

I for one am in favour of Joe's idea: for an unixian it sounded a bit
alien anyway to see the environment change in the middle of the life
of a process...

(Now it remains to be checked whether some popular extension will veto
this move because it spawns a process without using the Tcl API,
making it hard to apply this delay-update scheme)

-Alex

.

Relevant Pages

  • Re: Windows env-Bug?
    ... Controlling the process environment is one of the ... basic methods of the "interprocess communication", ...
    (comp.lang.tcl)
  • Re: should these be fixed for python 2.4?
    ... Alexander Schmolck wrote: ... spawn* functions but ... ... I'll add the above to my toolbox. ... environment, v = something, l = something else) but I ...
    (comp.lang.python)