Try not to over-complicate things; if its as simple as a single line of code
then please take the short period of time to do it.  Write a one-line test
that passes in DEBUG mode and fails in RELEASE mode and attach it to a JIRA
issue.

Clearly not *all* queries fail in RELEASE mode else nearly everyone running
NH today would probably have exceptions rather than query results so its
reasonable to assume that with your specific failure case in mind you could
more easily/quickly produce this failing test than a committer could
starting from absolute zero.  Your effort to provide a failing test (even if
extremely brief) in helping the NH team quickly isolate the problem reduces
the friction in our trying to isolate the issue.

Else, we could just as easily say "we wrote a query that *didn't* throw,
ergo issue is closed as cannot-repro", which probably wouldn't satisfy you
either.  At which point you would reply with still more info about your
issue and we would be right back here asking you for a failing test that
demonstrates the problem conclusively :)

Which is why we ask for a failing test that demonstrates your issue.  In a
project this size, the thrashing created by all the back + forth that is
otherwise required represents significant overhead that we need to find ways
to reduce.  "Code is worth a thousand words" and its not open to
interpretation, mis-communication, or any of the other things that slow down
effective communication about bugs/issues.

Hope this helps to clarify why we ask for a failing test from all reporters.

Regards,

Steve Bohlen
[email protected]
http://blog.unhandled-exceptions.com
http://twitter.com/sbohlen


On Tue, May 31, 2011 at 5:42 PM, Beefy <[email protected]> wrote:

> Well that's disheartening. I would assume that all of this get's
> tested in both Debug and Release mode, correct? Any simple test I
> could write would do nothing but the most basic Session.Query<T>()
> call. It would just have to be attempted when compiled as a Release
> mode build instead of a Debug mode build.
>
> There doesn't seem to be anything special I am really doing in this
> case.
>
> / Michael /
>
> On May 31, 5:11 pm, Fabio Maulo <[email protected]> wrote:
> > I sent you two links.
> > The first is where start.
> >
> > To create a failing test what you need is just your know-how about your
> > problem; you don't need a special know-how about NHibernate.
> >
> > If you like to have a open issue, with your name, for a bunch of months
> > before close it as "Incomplete" then you can file a new ticket with the
> same
> > information you sent here.
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > On Tue, May 31, 2011 at 6:06 PM, Beefy <[email protected]> wrote:
> > > Love to, but I have no idea where to even start for it, and I must bow
> > > to the experience of this group in this matter. I'm hoping someone
> > > with the know-how about the internal workings of NHibernate might pick
> > > up and run with this. I can only offer my observations, unfortunately.
> > > Would you like me to still open a ticket there?
> >
> > > / Michael /
> >
> > > On May 31, 4:54 pm, Fabio Maulo <[email protected]> wrote:
> > > > As usual a nice self explained test attached to a JIRA ticket (after
> you
> > > > have really realized that the test fail) will be more clear than
> hundreds
> > > of
> > > > words.
> > > > To create a test:
> > >http://nhforge.org/blogs/nhibernate/archive/2008/10/04/the-best-way-t.
> ..
> >
> > > > <
> http://nhforge.org/blogs/nhibernate/archive/2008/10/04/the-best-way-t..
> > > .>To
> > > > create a JIRA ticket:http://216.121.112.228/secure/Dashboard.jspa
> >
> > > > On Tue, May 31, 2011 at 5:46 PM, Beefy <[email protected]> wrote:
> > > > > Hmm... interesting, haven't used Single before. That allows me to
> get
> > > > > rid of a line of code. Thanks for that.
> >
> > > > > However, it still does not fix the underlying issue.
> >
> > > > > / Michael /
> >
> > > > > On May 31, 4:41 pm, Fabio Maulo <[email protected]> wrote:
> > > > > > > Not the same thing. I need to know that there is one, and
> exactly
> > > one,
> > > > > > > result.
> >
> > > > > > So: Single()
> > > > > > or SingleOrDefault() if you allow no-results.
> > > > > > --
> > > > > > Fabio Maulo
> >
> > > > --
> > > > Fabio Maulo
> >
> > --
> > Fabio Maulo
>

Reply via email to