|
|
Boost Testing : |
From: Beman Dawes (bdawes_at_[hidden])
Date: 2005-05-11 18:57:25
"Victor A. Wagner Jr." <vawjr_at_[hidden]> wrote in message
news:6.2.0.14.2.20050510172121.07891f20_at_mail.rudbek.com...
> At Wednesday 2005-05-04 05:25, Beman Dawes wrote:
>
>>"Victor A. Wagner Jr." <vawjr_at_[hidden]> wrote in message
>>news:6.2.0.14.2.20050503192915.08864920_at_mail.rudbek.com...
>> > Well, I found out why the tests are failing (kinda) ... this is run
>> > manually with outputting timestamps before and after in a batch file.
>> > The difference is 10min 48seconds
>> > I don't think the build_monitor waits that long.
>>
>>Not sure why the timestamps show that much difference. The apparent bug
>>this
>>test is detecting is network related, so it is barely possible network
>>issues are causing a long pause, but that seems very unlikely.
>
> you've not played with the networks on Windows enough, lol.
That's probably true:-)
>>Could you please do this:
>>
>>* Uncomment line 84 of operations_posix_windows.cpp so it reads #include
>><iostream>
>>* After line 355 in operations_posix_windows.cpp, which reads UINT err =
>>::GetLastError();, add a line:
>>
>> std::cout << "error " << err << '\n';
>>
>>* Compile and run operations_test, and let me know the error code being
>>returned just before the (125) error message above displays.
>
> done
>
Thanks!
> and the output
> T:\Projects\boost\bin\boost\libs\filesystem\test\operations_test.test\vc-7_1\debug\threading-multi>type
> operations_test.output
> Platform is Windows
> initial_path().string() is
> "T:/Projects/boost/libs/filesystem/test"
> initial_path().native_file_string() is
> "T:\Projects\boost\libs\filesystem\test"
> error 53
> error 2
> error 2
>
>...
>
> **** no errors detected
>
> EXIT STATUS: 0
Nuts! The test passed this time!
I see it is now passing on the posted regression results, too. Not sure what
to make of that. I did made a fix this morning, but it was to an unrelated
portion of the tests.
I'll keep an eye on the results for a few days to make sure it doesn't
regress.
Thanks,
--Beman