Are all tests supposed to pass?

May 22, 2013 at 4:32 PM
Edited May 22, 2013 at 4:33 PM
Hi,

Since I had some problems running the precompiled versions (WinDBG died when I loaded the extension) I decided to build it myself. I only have VS2012, but after the adjusting the python paths and installing boost everything builds. Some of the tests fail however:

1> creating typeInfo by the type name ... ERROR
1> testStaticField (typeinfo.TypeInfoTest) ... FAIL
1> testArrayFieldSlice (typedvar.TypedVarTest) ... ERROR
1> testArrayFieldSliceNegative (typedvar.TypedVarTest) ... ERROR
1> testStaticField (typedvar.TypedVarTest) ... ERROR
1> testSymbolNameAddress (mspdbtest.MsPdbTest)
1> Lookup symbol by name/address ...
1> Module: ole32
1> Start: 72540000 End: 7269c000 Size: 15c000
1> Image: C:\Windows\System32\ole32.dll
1> Symbols: export symbols
1> Timestamp: 4ce7b96f
1> Check Sum: 164a37
1>
1> ERROR
1> testLocalVariable (localstest.LocalVarsTest)
1> Start new process and test local variables ... ERROR
1> Start new process and track change symbols exception ... FAIL

How worried should I be given that all other tests pass?
Coordinator
May 23, 2013 at 9:37 AM
They are supposed :). But in fact some tests are failing

1)
testStaticField (typeinfo.TypeInfoTest) ... FAIL
testStaticField (typedvar.TypedVarTest) ... ERROR
It is surprising for me. This test fails on vs2008 and 2010 due the bug in the MS DIA. But I think this bug is fixed for VS2012

2)
testArrayFieldSlice (typedvar.TypedVarTest) ... ERROR
testArrayFieldSliceNegative (typedvar.TypedVarTest) ... ERROR
pykd 0.1.x has a python slice support, but we have broken it in the 0.2.x

I know, it is not good to have a failing tests. I'm sorry :).
May 23, 2013 at 11:47 AM
On Thursday, May 23, 2013, kernelnet wrote:

From: kernelnet

They are supposed :). But in fact some tests are failing

1)
testStaticField (typeinfo.TypeInfoTest) ... FAIL
testStaticField (typedvar.TypedVarTest) ... ERROR
It is surprising for me. This test fails on vs2008 and 2010 due the bug in the MS DIA. But I think this bug is fixed for VS2012

This was using vs2012 update 2.

I know, it is not good to have a failing tests. I'm sorry :).
I suppose. But more important for me is whether this is worse than the official builds which would mean that something broke due to the upgrade to vs2012 or my build environment (eg due to different boost/python versions).

I'll try it and see if it explodes...