[prev in list] [next in list] [prev in thread] [next in thread] 

List:       ntp-hackers
Subject:    Re: [ntp:hackers] refclock ONCORE question
From:       Harlan Stenn <stenn () ntp ! org>
Date:       2010-11-19 18:03:43
Message-ID: 201011191803.oAJI3h44079013 () stenn ! ntp ! org
[Download RAW message or body]

Reg wrote:

> At that point someone complained and asked that they go to the syslog
> file also.  Fine.  Nowhere in the documentation is the function
> msyslog ever mentioned, so I used syslog.  I still have no idea what
> msyslog does, since there is no documentation.

Noted, and I would be quite happy if (many) people started to do this.
It doesn't have to be "somebody else's job".  As an aside, along the way
people can start writing tests for functions as well.

> And to add a complaint I have made before, rather than adding
> additional 'features' to ntp, I would much prefer that the things that
> are there were documented.  Even a few lines of text at the beginning
> of a routine would help.  I know its a lot less sexy, and it wont
> happen, but I have found new variables (that were significant to the
> ONCORE driver) added to the code and just never mentioned.  Ditto with
> routines like msyslog that a driver should be using.  There should be
> a list of these things.

autogen can produce API documentation, and I gather doxygen is far less
painful to setup and use now than it was the last time I looked at it.

> My other complaint, as long as Im complaining, is that changes made to
> the Development version of ntp dont seem to ever make it into the
> Stable version, unless you put the change both places.  I put a change
> in the development version almost a year ago, and its still there.  If
> even the 'additions' to the interface are going to be inserted into
> the stable version, and not pass thru the development version for
> further testing, then what is the reasoning for having both versions?

This one is easy, and it's even documented:

 http://support.ntp.org/bin/view/Dev/MaintainerIssues#About_the_stable_and_dev_reposit

 http://support.ntp.org/bin/view/Dev/MaintainerIssues#Fixing_a_bug_in_both_stable_and

The bottom line is that it is up to the refclock maintainer to decide where
a bugfix/enhancement should be put.

As for when -dev will become -stable (ie, 4.2.8), we are pretty close to
beginning that release cycle.

In my opinion, the NTP effort is doing a pretty amazing job considering
the scarce resources we have available to us.

Take a look at http://freshmeat.net, at the "Most Vital Projects" list
on the RHS of the page, for example.

I'm spending a chunk of my time working on increasing the resources we
have, as it is clear to me that we can do significantly better with more
resources (financial, especially) for this critical (and pretty much
currently unfunded) piece of core internet infrastructure.

> OK, Ill step down off my soapbox.  The patch is coming, but my
> main machine here has developed some hardware problems, so I 
> wasnt sure I would even get thru this mail message.  It may be a
> few days before I track the problem down to the point that I can
> actually recompile ntp.

No worries, and I appreciate both your comments and feedback as well as
your ongoing efforts here.

H
_______________________________________________
hackers mailing list
hackers@lists.ntp.org
http://lists.ntp.org/listinfo/hackers
[prev in list] [next in list] [prev in thread] [next in thread]