Re: [eigen] Feeback to Rev #914777. (pre RC1)

[Gael Guennebaud <gael.guennebaud@xxxxxxxxx>]

hi,

it's all green now (http://my.cdash.org/buildSummary.php?buildid=7331)

for the internal error, I have no clue.

Gael.

On Fri, Jan 23, 2009 at 1:10 AM, FMDSPAM <fmdspam@xxxxxxxxx> wrote:
> Gael Guennebaud schrieb:
>>
>> I cannot reproduce that internal error. Here everything compile file
>> using MSVC 9.00.21022.08 with following flags:
>>
>> /DWIN32 /D_WINDOWS /W3 /Zm1000 /EHsc /GR /arch:SSE2
>>
>> (I also tried without /arch:SSE2)
>>
>> Note that I'm using nmake to call MSVC but that should not make any
>> difference.
>>
>>
>
> In my Setup (MSVC SP1, Vista etc, etc)
> right out of the repository Rev #915248 the test suite is compiling
>
> without /arch:SS2 without error! but four warnings (for the same), see
> below.
> with /arch:SS2, the internal compiler error 1001 (located by MSVC in the
> cachefriendlyproduct.h line 497) still occur.
>
> If you have any idea, what kind of tests I should do, please ask for.  I'm
> totally clueless.
> Is there anybody out there with MSVC Express Edition SP1, who is able to
> confirm my problems?
>
> without SSe2, all tests are passed, except the following:
>>>>>>>>>>>>>>>>>
> 18/38 Testing: commainitializer
> 18/38 Test: commainitializer
> Command: "C:/b/test/Release/test_commainitializer.exe"
> Directory: C:/b/test
> "commainitializer" start time: Jan 23 00:43 Mitteleurop&lt;-28&gt;ische Zeit
> Output:
> ----------------------------------------------------------
> Initializing random number generator with seed 1232667781
>
> Repeating each test 10 times
>
>
>
> This application has requested the Runtime to terminate it in an unusual
> way.
> Please contact the application's support team for more information.
>
> ----------------------------------------------------------
> "commainitializer" end time: Jan 23 00:43 Mitteleurop&lt;-28&gt;ische Zeit
> "commainitializer" time elapsed: 00:00:08
> ----------------------------------------------------------
> <<<<<<<<<<<<<<<<<<
>
>
> Four Warnings for the same thing in submatrices.cpp Line 144, 145, 147, 148:
> warning C4244: 'argument' : conversion from 'int' to 'const
> std::complex<float>::_Ty', possible loss of data
> in the expression "ones.col(c1).sum() == Scalar(rows)" with type of Scalar
> is std::complex<float>
> by CALL_SUBTEST( submatrices(MatrixXcf(3, 3)) );
>
>
> Frank
>
>
>
>