Miguel de Icaza wrote:
Well, I think I hit the wrong 'scenario' (see scenario 3 below).Should I be able to run the tests on Windows? Or does it only run ok in Linux (and also: does it run with mono (windows/linux), MS .NET - this is interesting for me to know)
Yes, the tests should run on Linux and Unix. What exactly does fail?
We have the following variables:
* compiler (used to build test) : mcs (mono), csc (microsoft)
* referenced assemblies : mono, microsoft
* runtime (used to execute test): mono, microsoft .net
* etc
This leads to several scenarios (not all mentioned here):
1) compile test with mcs, referencing mono assemblies, running in mono
..
3) compile test with csc, referencing mono assemblies, running in microsoft .net
..
5) compile test with csc, referencing microsoft assemblies, running in microsoft .net
..
etc.
Now the NUnitGuidelines document states : "Test your test with the microsoft runtime". What does this mean?
I think this means scenario 5 should be possible (and should be easy doable using the existing makefiles/.build files).
If it's 'difficult' to fullfill this, people probably won't do it.
If you're running linux, scenario 1 is probably the most used one (and thus most tested).
The problem is that libxslt from GNOME can only handle files or
DOMs that it has loaded itself. So the trick is supporting the
DOM and XPathNavigator interfaces; What we do is we dump those
into a temporary file, and then handle the temporary file to
Xslt.
This works (but might need some tests).
Ok, I think I will start with test for the existing items first. Is there a Test guideline on how to deal with input/output files? --- Jeroen Janssen
_______________________________________________ Mono-list maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/mono-list
