gpsd-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [gpsd-dev] Hal's sirf1 bug fixed


From: Eric S. Raymond
Subject: Re: [gpsd-dev] Hal's sirf1 bug fixed
Date: Sat, 24 Jan 2015 11:29:39 -0500
User-agent: Mutt/1.5.23 (2014-03-12)

Greg Troxel <address@hidden>:
> Is it a fair guess that WRITE_PAD is more likely to fix problems in the
> middle and CLOSE_DELAY missing trailing output?

I think so.  But the behavior is sufficiently random that almost anytbing
I think I know about it could be mere pareidolia.

> Is there a notion of blocking on master-side pty writes when the buffer
> becomes full?  I would think that there must be, although usually it's
> the other direction that blocks.   It seems like we should catch
> EWOULDBLOCK, or actually block.  So this is a bit mysterious to me.

Tell me about it!  I've been seeing various manifestations of this
problem since I first wrote gpsfake in 2005.

It used to be worse. I added some code to set a linger option on client 
sockets and gracefully shut them down when gpsd is signalled to terminate.
That reduced the problem.  I think - I'm not even certain of that.

Googling found some lore I didn't know here:

http://blog.netherlabs.nl/articles/2009/01/18/the-ultimate-so_linger-page-or-why-is-my-tcp-not-reliable

I may experiment with his techniques.
-- 
                <a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>

Attachment: signature.asc
Description: Digital signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]