Re: prompt_hook

Nope--haven't profiled it which is why I haven't qualified how much of a
performance hit it is.  I'm just removing things that probably weren't
architected well and didn't meet the needs I originally intended them for.

Also, I'm not entirely sure how to go about re-implementing it using the
telnetlib.  So one possibility is that I could check in the changes I've
made and we can work through re-implementing the pieces as we figure out
how to re-implement them.  I don't use prompt detection, so it's not
something that I can test easily.

/will


On Tue, 20 Apr 2004, Eugene wrote:

> On Tue, 20 Apr 2004 08:22:25 -0500 (CDT) will guaraldi
> <[email protected]> wrote:
> >
> >Here's why I'm asking: I just reimplemented the net module to use the
> >telnetlib (which is broken in regards to telnet options in previous
> >versions of Python [though I can't remember which ones--I think 2.2 but
> >it might have been 2.1--I keep meaning to look this up]) and in doing
> >so dumped all our prompt detection stuff.
> >
> >I was originally thinking it'd be useful for things like mappers and
> >anything else that really needs to link user commands with mud
> >responses. Josh implemented his commandresponse module at a higher
> >level and it seems to work a lot better.  I was never able to get my
> >commandresponse module (which worked with the prompt detection) to
> >work.
> >
> >So, given that I couldn't use it for what I was hoping to use it for,
> >I'm thinking it slows things down (running a regexp on all incoming
> >data...)
>
> It is another issue. I said I don't use the hook itself, but I'd like
> the detection stuff (including the regexp) to remain. I anyway detect a
> prompts by checking the incoming line with line.endswith("\n"), so I
> think it is convenient to have a prompt going through filters as
> separate line.  (I rely on prompts-as-separate-lines in another my
> modules too.)
>
> Generally, I think prompts should be considered as separate lines
> themselves, not as nondelimited part of some another line which is have
> to come later. And it cannot be done without the regexp checking.
>
> By the way, I don't think the regexp checking eats that many of CPU; did
> you perform a profiling? The regexp are usually pretty fast.


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click