Validation failures under Windows

[Simon Marlow]

Mitchell, Neil wrote:
>
> Thanks Thorkil, more information:
> 
> =====> 2228(normal)
> cd ./ghc-e/should_run && $MAKE --no-print-directory -s 2228
> </dev/null >2228.run.stdout 2>2228.run.stderr
> =====> 2636(normal)
> cd ./ghc-e/should_run && $MAKE --no-print-directory -s 2636
> </dev/null >2636.run.stdout 2>2636.run.stderr
> Actual stdout output differs from expected:
> --- ./ghc-e/should_run\2228.stdout.normalised	2008-10-17
> 16:03:32.386896400 +0100
> +++ ./ghc-e/should_run\2228.run.stdout.normalised	2008-10-17
> 16:03:32.386896400 +0100
> @@ -1,2 +1,2 @@
> -BlockBuffering Nothing
> +LineBuffering
>  BlockBuffering Nothing
> *** unexpected failure for 2228(normal)
> ====> Running ./ghci/prog001/prog001.T
> =====> prog001(ghci)
> cd ./ghci/prog001 && HC='c:/ghc-build/ghc/ghc/stage2-inplace/ghc'
> HC_OPTS='-dcore-lint -dcmm-lint -Di386_unknown_mingw32
> -dno-debug-output ' 'c:/ghc-build/ghc/ghc/stage2-inplace/ghc'
> --interactive -v0 -ignore-dot-ghci -dcore-lint -dcmm-lint
> -Di386_unknown_mingw32  -dno-debug-output     <prog001.script
>> prog001.run.stdout 2>prog001.run.stderr
> ====> Running ./ghci/prog002/prog002.T
> 
> Could it be that buffering has a different default on Windows?

2228 is now an expected failure on Windows.  See #2628.

> =====> break009(ghci)
> cd ./ghci.debugger/scripts &&
> HC='c:/ghc-build/ghc/ghc/stage2-inplace/ghc' HC_OPTS='-dcore-lint
> -dcmm-lint -Di386_unknown_mingw32  -dno-debug-output '
> 'c:/ghc-build/ghc/ghc/stage2-inplace/ghc' --interactive -v0
> -ignore-dot-ghci -dcore-lint -dcmm-lint -Di386_unknown_mingw32
> -dno-debug-output  -ignore-dot-ghci   <break009.script
>> break009.run.stdout 2>break009.run.stderr
> --- ./ghci.debugger/scripts\break008.stdout.normalised	2008-10-17
> 16:04:39.247127200 +0100
> +++ ./ghci.debugger/scripts\break008.run.stdout.normalised
> 2008-10-17 16:04:39.247127200 +0100
> @@ -1,3 +1,3 @@
> -Breakpoint 0 activated at ../Test3.hs:1:13-14
> -Stopped at ../Test3.hs:1:13-14
> -_result :: [a] = _
> +Cannot find default module for breakpoint.
> +Perhaps no modules are loaded for debugging?
> +[]
> *** unexpected failure for break008(ghci)
> 
> And a GHCi debugger issue.

I haven't seen this, but I have noticed that one or two of the GHCi 
debugger tests seem to be failing on Windows, I'll try to look into it shortly.

> As a secondary issue, it appears that the failure reports are occuring
> after the script has started to print out information about the next
> test.

That's probably because the testsuite is running in parallel, which it does 
automatically when validating if you have a new enough version of Python.

Cheers,
	Simon