[Nut-upsdev] SSL test regression

[Arnaud Quette]

2010/2/17 Arjen de Korte:
> Citeren Charles Lepple:
>
>> I wonder if there are any cross-compilation targets we could use to test
>> some of the word-size assumptions. Also, we could add in some static
>> analysis tools.

FYI, I submitted NUT to the Coverity Scan program
(http://scan.coverity.com) last september.
3 pings later, I still have no news, nor at least a first contact!
maybe some of you may want to second that request?

> Regarding the latter, I tried running 'splint' on a couple of source files a
> while back. It didn't take long to realize that this might indeed improve
> the quality of the code. But this will be a monumental task to do this NUT
> wide, since without added markup (to tell splint what we're doing), this
> will result in an endless list of warning messages.
>
> After nut-2.4.2 is out, I'll commit a couple of examples to show what would
> be needed.

excellent.

don't hesitate to complete the QA document:
http://buildbot.ghz.cc/~buildbot/docs/r2351-444/website/nut-qa.html
(generated from docs/nut-qa.txt)

I would really like to see the static code analysis subject addressed.

Arjen, have you compared / tried other systems like Frama-C, BLAST or Clang?

before investing in one system or another (which is part of 2.8), I
would like to have a clear picture of the situation.

cheers,
Arnaud