Subject: Re: [ecasound] edi-28 - first implementation
From: Kai Vehmanen (k_AT_eca.cx)
Date: Wed Oct 16 2002 - 19:04:21 EEST
On Wed, 2 Oct 2002, Tony Leake wrote:
>>I've just committed the first implementation of edi-28, "Extending the
>>EIAM command set to allow writing standalone ECI implementations", to
> I have just discovered that I only have one way communication to a forked
> process in PHP, so I won't be able to do anything with this untill thats
> fixed. I was sure there was a workaround but it doesn't work.
Yup, that is a common problem. Unfortunately the popen() C-lib function,
which is the easiest way to fork+control a child process, only supports
one-way communication. I, too, had to code around this in libecasoundc_sa
using fork+exec+2xpipe()... Same thing probably needs to be done for the
PHP module for popen(). I wouldn't be suprised if python implementation
had the same limitation. ;)
> I have been playing with this in interactive mode though and it seams all
> very straightforward. I was a little confused at first by the messages with
> different log levels, but I'm sure if I had 2 way communications this could
> be implemented in a couple of hours.
Yup, it seems to work quite well. I've written a small testsuite
(ecasound/libecasoundc/libecasoundc_tester.c - you can compile and run it
with 'make check') and most things work ok. There's still something wrong
in the list-of-strings handling, but fortunately I now have test-cases
to fix+verify the problem. And of course, now that we have ecasound.el,
there's already two independent implementations of the ECI-standalone!
-- http://www.eca.cx Audio software for Linux!
This archive was generated by hypermail 2b28 : Wed Oct 16 2002 - 19:02:54 EEST