[Nut-upsdev] Driver for Powercom BNT/KIN/IMP/IMD series

[Charles Lepple]

On Dec 24, 2007 9:04 AM, Alexey Sidorov <alex at reutman.ru> wrote:
> Only for BNT/KIN/IMP/IMD:
> power and linevoltage whould detect automatically now
> also you don't need any other coefficients in ups.conf
> added shutdown.stayoff and shutdown.return commands
> start battery.test on driver start
>
>  Tested on: BNT-1200AP

Since this has only been tested on one model so far, we should
probably apply it to the SVN trunk first, and backport it to Testing
(which is the branch where 2.2.0 and 2.2.1 came from) after it has
been tested on some other models, or has been proven to be
non-invasive.

> Example of ups.conf:
> [<upsname>]
> driver = powercom
> port = /dev/ttyS0 # or which port you use
> type = BNT          # or KIN or IMP (for IMP/IMD) or Trust or Egys or KP625AP,  your UPS model
>
> Also, i think Trust and Egys - it's just KIN by another manufacturer, isn't it?
> Also, KP625AP - what is it? What full name of this model?

It looks like you are adding some possible values for the options -
would you mind updating the documentation for the driver
(man/powercom.8)?

In a similar vein, it's a good idea to write up a summary of your
changes in the same format as the ChangeLog file. This is what helps
us collect the top-level changes that users need to know when they are
upgrading NUT.

Have you been in contact with the original author of this driver? If
you are planning to make other changes to the driver, the simplest
thing might be for you to check out a copy of the trunk from SVN, and
send the output of "svn diff" (which is like the output of "diff -u")
for easy identification of the changes.

It looks like you changed the name of the nut_shutdown() function back
to "shutdown". The nut_shutdown() name is intentional - there is a C
library function called shutdown() that is used for sockets.

We have convenience functions for setting the flow control pins now:
http://boxster.ghz.cc/projects/nut/changeset/981 . You might be able
to apply that patch (see "Unified diff" link at the bottom of that
page) with a large fuzz factor to your copy of the code, but using SVN
would probably be easier.

-- 
- Charles Lepple