[regression][winapi] Tests not being run
Hi, I wrote about it before, but now the problem becomes more weird. The problem is that the tests for Boost.WinAPI are not being run for some reason. What's weird is that now there is one test in the matrix (http://www.boost.org/development/tests/develop/developer/winapi.html) that isn't part of the Boost.WinAPI test suite. I assume that entry is part of the formal checks for the library conformance, since it is present for all libraries. Could someone please have a look why the tests for Boost.WinAPI are not being executed? PS: The link for the __boost_check_library__ entry is broken: https://github.com/boostorg/winapi/blob/developstatus/boost_check_library.py I believe, there's a slash missing between develop and status, and also the repository name should be boost rather than winapi.
On Mon, Jul 11, 2016 at 6:27 AM, Andrey Semashev
Hi,
I wrote about it before, but now the problem becomes more weird.
The problem is that the tests for Boost.WinAPI are not being run for some reason. What's weird is that now there is one test in the matrix ( http://www.boost.org/development/tests/develop/developer/winapi.html) that isn't part of the Boost.WinAPI test suite. I assume that entry is part of the formal checks for the library conformance, since it is present for all libraries.
Could someone please have a look why the tests for Boost.WinAPI are not being executed?
First.. The test are being run. Jut not getting reported. Second.. Are you sure they used to show up? I'm seeing this as the reported test info: boost-test(COMPILE) "winapi/~hdr-decl/interlocked~hpp" : "libs/winapi/test/compile/decl_header.cpp" Which seems to match this code: # All '/' are replaced with '-' because apparently test scripts have a problem with test names containing slashes. local test_name = [ regex.replace $(rel_file) "/" "-" ] ; local decl_test_name = ~hdr-decl/$(test_name) ; local use_winh_test_name = ~hdr-use-winh/$(test_name) ; local pre_winh_test_name = ~hdr-pre-winh/$(test_name) ; local post_winh_test_name = ~hdr-post-winh/$(test_name) ; But the code seems to contradict itself by adding "/" in the test names after saying it can't have "/". If this used to work.. I suspect my addition of the global library check test has illuminated this rather annoying limitation in the test reporting. But one easy way to fix it is to remove the additional "/" you are adding in the *_winh_test_name vars. -- -- Rene Rivera -- Grafik - Don't Assume Anything -- Robot Dreams - http://robot-dreams.net -- rrivera/acm.org (msn) - grafikrobot/aim,yahoo,skype,efnet,gmail
On 07/11/16 15:26, Rene Rivera wrote:
On Mon, Jul 11, 2016 at 6:27 AM, Andrey Semashev
wrote: Hi,
I wrote about it before, but now the problem becomes more weird.
The problem is that the tests for Boost.WinAPI are not being run for some reason. What's weird is that now there is one test in the matrix ( http://www.boost.org/development/tests/develop/developer/winapi.html) that isn't part of the Boost.WinAPI test suite. I assume that entry is part of the formal checks for the library conformance, since it is present for all libraries.
Could someone please have a look why the tests for Boost.WinAPI are not being executed?
First.. The test are being run. Jut not getting reported.
Second.. Are you sure they used to show up?
No, they never showed up in the matrix but they ran fine locally last time I checked.
I'm seeing this as the reported test info:
boost-test(COMPILE) "winapi/~hdr-decl/interlocked~hpp" : "libs/winapi/test/compile/decl_header.cpp"
Which seems to match this code:
# All '/' are replaced with '-' because apparently test scripts have a problem with test names containing slashes. local test_name = [ regex.replace $(rel_file) "/" "-" ] ; local decl_test_name = ~hdr-decl/$(test_name) ; local use_winh_test_name = ~hdr-use-winh/$(test_name) ; local pre_winh_test_name = ~hdr-pre-winh/$(test_name) ; local post_winh_test_name = ~hdr-post-winh/$(test_name) ;
But the code seems to contradict itself by adding "/" in the test names after saying it can't have "/".
Hmm, somehow I haven't noticed it earlier, thanks for pointing this out. I'll remove the slashes and see if that helps.
participants (2)
-
Andrey Semashev
-
Rene Rivera