wasm-demo/demo/ermis-f/python_m/cur/1553

45 lines
1.6 KiB
Plaintext

From: tim_one at email.msn.com (Tim Peters)
Date: Sun, 25 Apr 1999 17:29:31 -0400
Subject: help with os.popen()
In-Reply-To: <87yajhxm52.fsf@psyche.evansnet>
Message-ID: <000001be8f62$b5494f00$669e2299@tim>
Content-Length: 1388
X-UID: 1553
[Tim, wrt popen() on Win32]
> I'll figure this all out if I ever get a free year <0.7 wink>.
[Carey Evans]
> Why bother? The Tcl people have done it already, I've had very few
> problems with pipes there, and it's under a generous license. There's
> some interesting comments in tclWinPipe.c and stub16.c too, and an
> amazing number of special cases.
>
> I'd be eternally grateful[1] if pipes worked as well in Python as they
> do in Tcl, although I suppose I could try rewriting tclWinPipe.c using
> the Win32 extensions. Actually, it would be quite nice if I could
> just get the return value from os.system() or os.spawnv() like I can
> from Tcl's "exec".
>
> [1] I'd buy the author a drink if I met them, anyway.
So, Carey, this particular problem bothers you, and you believe you know how
to fix it. How about buying yourself a drink <0.9 wink>?
After staring at the ~2,700 lines of Tcl source you referenced, my reaction
remains "I'll figure this all out if I ever get a free year".
Hip's original Perl-for-Win32 port shipped its own (non-interactive) 32-bit
shell, so that Perl popen/backtick/system constructs stood some chance of
working. The latest Perl-for-Win32 source appears to have dropped that,
though, with scattered vague recommendations that you install your own
non-brain-dead PERL5SHELL (4DOS/NT?).
about-to-boost-it-to-two-years<wink>-ly y'rs - tim