Re: [fluid-dev] Implementing the TCP command shell on Windows

[Josh Green]

On Tue, 2003-10-07 at 17:29, Ken Ellinwood wrote:
> I've been thinking about implementing the tcp command server for Windows
> recently and I'll have a few days to actually do it later this month.
> 
> Its kind of ironic that the TCP command shell doesnt work on Windows, because
> the primary reason for its existence is that you cant read and write to the
> stdout/stdin of a subprocess on this platform.
> 

Hmm, I'm not familiar with Windows programming myself and don't have
access to a machine to test things on. Anyone else want to fix this?
I'll add it to the TODO list, but I doubt I'll get to it anytime soon.
I've been rather busy lately with work and haven't had a lot of time for
FluidSynth.

> Anyway, the FluidGUI application not only uses the tcp command shell on Linux,
> but it also read/writes to the stdio of the subprocess in order to get the
> errors and warnings produced by fluidsynth.  
> 
> I propose that when the options -i and -s are given to fluidsynth (meaning 
> that
> the only way to send/receive output is to use the tcp shell server) that
> messages be redirected to the tcp port prepended with "error: ", "warning: ",
> "info: ", etc., so that a front-end application can distinguish between normal
> output from commands and the errors, warnings and informative messages that 
> the
> synth also produces.
> 

Good idea. I'll put that on the TODO as well.

> Ken
> 

Cheers.
        Josh Green