[Mono-dev] [Fwd:[Mono-patches] r56597- trunk/mcs/class/System.XML/Test/System.Xml.Schema]

[Gert Driesen]

> -----Original Message-----
> From: mono-devel-list-bounces at lists.ximian.com 
> [mailto:mono-devel-list-bounces at lists.ximian.com] On Behalf 
> Of Atsushi Eno
> Sent: dinsdag 7 februari 2006 4:55
> To: Gert Driesen
> Cc: 'mono-devel mailing list'
> Subject: Re: [Mono-dev] [Fwd:[Mono-patches] r56597- 
> trunk/mcs/class/System.XML/Test/System.Xml.Schema]
> 
> Hi,
> 
> >> Please revert this patch which is done without any agreement.
> > 
> > I know, but I thought it was a very trivial patch.
> 
> Well, my bad, agreement part does not that matter (usually
> I am regarded as to have agreed to commit such patches just
> because I am the maintainer, so I put status for some patches
> as a disclaimer). If you'll take future responsibilities from
> the change then it is no problem though.
> 
> >> There is a reason I marked it as NotWorking - it is for
> >> compatibility kids. 
> > 
> > Why mark something "NotWorking" if it is indeed working ?
> 
> Umm, you are right on the different aspect. It should be
> indeed [Ignore] since the test itself is wrong, as well as
> our code is not correct (in compatibility aspect).

I'm not sure I understand this. The test passes on both Mono (1.0 and 2.0
profile), and MS.NET 1.x after my minor change, then why not just mark the
test NotDotNet on the 2.0 profile. That will at least ensure the Mono
implementation keeps matching the MS.NET 1.1 implementation.

> NotDotNet should be used only in such case that 1)MS has buggy
> implementation, or 2)the test is Mono dependent, or so.

Yes, I agree.

> >> I am not likely to fix this trivial stuff
> >> soon, but someone might notice the impl. difference from it.
> >> Marking NotDotNet is clearly wrong: we do not test and complain
> >> if mono fails to assert this behavior.
> >>
> >> (Leaving Assertion->Assert changes is OK.)
> >>
> >> Actually why did you mark it as NotDotNet? It does not fail under
> >> .NET 2.0 (or do you have different .NET 2.0 than mine, 2.0.50727 ?)
> > 
> > On my systems, it only failed on .NET 2.0 (but worked fine 
> on Mono 1.0/2.0
> > and MS.NET 1.x).
> 
> OK, your change on #4 was wrong: it changes AssertNull() to 
> Assert.Is*Not*Null() which obviously changes test semantics.

Yes, but that change allows us to enable the test on Mono (1.0 and 2.0
profile) and MS.NET 1.x, which is far more useful than a test that is never
executed (because its marked NotWorking or Ignore).

> I'll revert the change and remark it as Ignore. No need to waste
> extra time.

I really don't want to waste time for either of us, just want to avoid these
discussions in the future by making sure I understand you.

Sorry for the fuzz

Gert

PS. Have you looked at my other small XmlSchema patch (bug #77491) ? Also
you attached patches for bugs #66031 and #75831, and change they will be
committed soon ?