Re: [dotNetRDF-Develop] dotNetRDF patch for Stardog 2.2

[Ron M. Z. <ro...@ze...>]

Rob, I'll take a look through the unit tests, although I can't promise
anything. I hate to say it but I've never used NUnit or done any unit
testing so I'll have to get up to speed on that a bit. Although looking
through StardogTests.cs, it looks simple enough: I assume a Stardog
instance is launched (manually? automatically?) before the tests start, and
it just connects and runs through anything marked as a Test?

It even looks like there already is one test for reasoning:
StorageStardogReasoningQL? It doesn't test for all the variations of
reasoning that Stardog supports but it would have detected this bug once
you tested against Stardog 2.2.

Ron



On Mon, Jul 7, 2014 at 6:44 AM, Rob Vesse <rv...@do...> wrote:

> Ron
>
> I took a quick look over the changes and they looked fine to me, thanks
> for debugging and fixing this
>
> Btw if you have any time could you attempt to create some relevant unit
> tests for this that we could add to the stardog tests so this is more
> easily caught If anything changes in this regard in the future?
>
> Cheers,
>
> Rob
>
> From: Ron Michael Zettlemoyer <ro...@ze...>
> Date: Saturday, 5 July 2014 04:39
> To: Rob Vesse <rv...@do...>
> Subject: dotNetRDF patch for Stardog 2.2
>
> Hey Rob,
>
> As you might have seen in the back & forth with Clark & Parsia on Twitter,
> they changed how the reasoning mode is set for Stardog. The deprecated
> using the HTTP headers back in 2.0 and finally disabled it in 2.2; it's
> been replaced with a URL parameter. So I found that reasoning was broken
> when using dotNetRDF and Stardog 2.2. I checked in some minor code changes
> that will get it working again. I left the old headers in place because
> Mike said that only 2.2+ support the new URL parameter. I tested the change
> against Stardog 2.1.2 and 2.2 and it worked as expected.
>
> It's been awhile since I've been at all active in the project so hopefully
> I checked it into the right place.
>
> Ron
>
>