[asterisk-dev] Quick question about this bug report

[John Lange]

If the astcli is a new feature and it is designed to replace accessing
the console from scripts; shouldn't the changes to the console that
cause script access to not work well have been saved until astcli is
released in 1.6?

Don't get me wrong, now that we've changed to using astcli under 1.4 I'm
much happier with the performance of our scripts but I'm worried that
there are others out there that have a lot of scripts that rely on
console access who can't easily fix the problem.

-- 
John Lange
www.johnlange.ca


On Fri, 2008-06-13 at 16:25 -0500, Tilghman Lesher wrote:
> On Friday 13 June 2008 15:53:50 John Lange wrote:
> > I reported this bug about console delays that was closed:
> >
> > http://bugs.digium.com/view.php?id=12851
> >
> > Corydon mentions the script atscli as a solution. I just want to mention
> > that as of 1.4.21, there is no astcli script in contrib.
> >
> > I assume that is an oversight because I was able to locate it in
> > svn/trunk and it works great.
> 
> It's technically a new feature.
> 
> > And just a minor comment; though the console isn't supposed to be suited
> > for scripting I still think it is exhibiting some strange behaviour
> > that's worth debugging. Why is it hanging after -rx commands? It never
> > used to do that. It's a small delay but it may be hinting at a larger
> > problem.
> 
> The way that the -rx command works is kind of a hack, and due to output
> delays, it's always been possible for output to be truncated.  By lengthening
> the time that the remote console waits for output, it is more assured that the
> output of the command is complete.  Specifically, -rx is exactly the same as a
> remote console, except that it waits for output only for a certain amount of
> time and then disconnects.  The delay was simply lengthened from 100ms to
> 500ms, to ensure that all output had the chance to make it through the pipe
> to the output device.
> 
> Another annoyance that we fixed recently was that on fairly busy servers, -rx
> would never disconnect, because the console did not distinguish between
> verbose console output and command output.  This has been changed, such
> that all verbose output is prefixed by a special preamble, which is then
> detected and removed at display time.  This also means that the output of
> -rx is now much cleaner than it has been in the past.
>