[Mono-devel-list] Re: valid-sa-100 testcase

[Atsushi Eno]

Hi Eyal,

> Just to set things straight. The patch was rejected since it was incomplete
> caused regressions and you didn't like the approach. The lack of concrete
> test case at the time that motivated the patch prevented us from continuing
> the discussion at that time. Now that the lost test was found :-) we are
> continuing the discussion.

That is also true to me that I am not motivated for that problem.
MS.NET is not also doing well on entity processing and I found no
serious existing solutions, so I just kept away from it.

> Now back to the test, how do you propose to fix the issue? What is your
> analysis of the issue?

Honestly, entity handling stuff is so minor that even MS.NET does
not work fine (for example RDF DTD that Joshua posted could not
be read fine), so am not fixing it soon. But of course I'd
appreciate if you guys or anyone would fix the problem :-)

> Regarding XmlDocument use, how do you think we should test it? 
 > When you changed xmltest.cs did you have cases where XmlDocument
 > didn't work in Mono but worked in MS?

I don't remember even whether I was aware of such differences or not.
I was not always paying attention to MS.NET behavior (since there
are many tests that .NET does not work fine). There are some cases
that (at least) XmlTextReader.Read() and XmlDocument.Load() differ
in rejecting input xml or not e.g. XmlDocument.Load() rejects
undeclared entities while XmlTextReader does not.

Atsushi Eno