Hello Charles and Eric,

The windows_port branch on my repo is mostly a rebase on latest master. I still 
have some work on this before pushing and of course not enough time for it. 
Anyway the branch on the official repo should be functional.
But I guess the main issue if you want to reuse the windows code is that it has 
been written to be compiled in a MinGW environment, not with MS building tools. 
So you will probably have to modify some NUT code anyway.

Regards,
Fred




________________________________

________________________________


From: Nut-upsdev 
[mailto:nut-upsdev-bounces+fredericbohe=eaton....@lists.alioth.debian.org] On 
Behalf Of Charles Lepple
Sent: jeudi 17 octobre 2013 02:01
To: eric kreuwels
Cc: nut-upsdev list; Frederic BOHE
Subject: Re: [Nut-upsdev] UPSC based Windows Client

[Please remember to copy the list.]

On Oct 15, 2013, at 8:45 AM, eric kreuwels wrote:


Hi Charles

Thanks for your answer. Good questions/suggestions. I copied both the WinUPSC 
and the installer project on my GDRIVE and shared it with you:
https://docs.google.com/file/d/0B0mSjYYj84-6ZGM4MWZfMnFxbUU/edit?usp=sharing

Archiving WinUPSC in the NUT GitHub would be my preference. Delivery as part of 
te Windows distribution is basically a separate decision for you guys.

Steps I have in mind:

  *   Check how WinUPSC will is perceived by the other developers! Then decide 
to setup a project in NUT GitHub
  *   Maybe a good idea to have a code review:

     *   Especially how I modified the UPSC code for polling. For example how 
standardized is the parameter list? I made assumptions here.
I haven't diffed your code against either NUT or WinNUT, but after a quick 
check, I'm not sure I understand your question here. If you are referring to 
the UPS variables, there is no guarantee that a specific variable will be 
available (except "ups.status"; think of a contact-closure UPS which can only 
report OL/OB/LB).

As an aside, in upsc.c, you use sprintf() with strlen(). There is a 
snprintfcat() utility in common/common.c

     *   I tried not to violate any licence and only re-use freeware, but don't 
fully comprehend these licences. Can these licences collide?
They can, potentially. The one thing I spot-checked was the icon page, and I 
didn't see any license information there.

     *   Are the credentials in the about box and the manual are sufficient :-)
I'm not sure where the About box is defined. The manual looks okay, although I 
wonder if the app code falls under the blanket "all rights reserved" clause: 
http://www.codeproject.com/info/TermsOfUse.aspx

It probably comes down to whether this code is boilerplate for all win32 
applications with an icon. I'm not the right person to make that call.

  *   Fix blocking issues for a first launch. My testing setup is limited (I 
only have Eaton UPSes). Feedback from others is really needed.
One way to test your code is to use the dummy-ups driver with the sample data 
files provided. data/evolution500.seq simulates an MGE/Eaton Evolution 500 
going on battery, then to low battery, then back online.

Of course, feel free to solicit testers with actual hardware. But with 
dummy-ups, you can basically simulate various combinations of variables, and 
there are many different examples in the list archives (search for "HCL"). 
Here's more information on dummy-ups: 
http://www.networkupstools.org/docs/man/dummy-ups.html

  *   Code cleanup; I like the idead to reference original NUT code (remove the 
duplicates in my project. I re-used the ported NUT files from WinNUT)
I CC'd Frederic Bohe who worked on the NUT windows_port branch. Perhaps he has 
some suggestions. The windows_port in his GitHub repo is different than the 
original one we converted from SVN, so I don't know the current status of that 
porting effort.

Looking forward to work with the development team

Eric Kreuwels

--
Charles Lepple
clepple@gmail



_______________________________________________
Nut-upsdev mailing list
[email protected]
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsdev

Reply via email to