[cfe-dev] [libcxx/test] How many tests should test it be passing?
Howard Hinnant
hhinnant at apple.com
Tue Feb 7 22:16:12 PST 2012
On Feb 7, 2012, at 6:20 PM, Matthew Abbott wrote:
> Hi all,
>
> I've noticed that test/testit is failing many more tests than the project's status page would have me believe. Here's the result I'm getting:
>
> ****************************************************
> Results for /Users/matt/libcxx/test:
> using clang version 3.0 (tags/RELEASE_30/final)
> Target: x86_64-apple-darwin11.2.0
> Thread model: posix
> with -std=c++11 -stdlib=libc++
> ----------------------------------------------------
> sections without tests : 1
> sections with failures : 545
> sections without failures: 518
> + ----
> total number of sections : 1064
> ----------------------------------------------------
> number of tests failed : 2618
> number of tests passed : 1706
> + ----
> total number of tests : 4324
> ****************************************************
>
> Should I be failing 545 sections still? I've tried a few different combinations of options, including the flags "-U__STRICT_ANSI__ -D__STDC_FORMAT_MACROS" and changing between c++11 and c++0x, but am getting the same number of tests failed regardless.
>
> I'm running on 10.7, and as you can see with clang-3.0 (the latest version available on macports, although I had been previously pulling the clang release from SVN and getting the same results). I thought I'd sanity check with the list, to see whether this failure rate is as high as I feel it is. I can give a more detailed output from the script if anyone desires, but it's a lot of template barf...
I would expect on the order of 100 test failures. Not 2000. But nothing in <atomic> is working. Here's a recent result for me:
****************************************************
Results for /Users/hhinnant/Development/temp_libcxx/test:
using Apple clang version 4.0 (tags/Apple/clang-418.0.2) (based on LLVM 3.1svn)
Target: x86_64-apple-darwin11.2.0
Thread model: posix
with -std=c++0x -stdlib=libc++
----------------------------------------------------
sections without tests : 1
sections with failures : 18
sections without failures: 1045
+ ----
total number of sections : 1064
----------------------------------------------------
number of tests failed : 64
number of tests passed : 4260
+ ----
total number of tests : 4324
****************************************************
It's a more recent compiler but I would not have expected that to make that much of a difference.
Howard
More information about the cfe-dev
mailing list