Boost logo

Boost Testing :

Subject: [Boost-testing] Unhelpful output in the test matrix
From: Marshall Clow (mclow.lists_at_[hidden])
Date: 2019-03-30 19:19:06


If you look at my testers, there are some failures that have very unhelpful
output.

For example:

https://www.boost.org/development/tests/develop/developer/Marshall-UBSAN-beast-clang-darwin-ubsan-_detail_base64-variants_.html
(which shows as a failure on the main 'develop' page)

Shows four links (for C++11, 14, 17, and 2a). So far so good.
Clinking on one (the C++11 -
https://www.boost.org/development/tests/develop/output/Marshall-UBSAN-boost-bin-v2-libs-beast-test-beast-core-_detail_base64-test-clang-darwin-ubsan-debug-cxxstd-11-iso-threadapi-pthread-threading-multi-visibility-hidden.html)
link, gets me this page:

Test output: Marshall-UBSAN - beast - _detail_base64 / clang-darwin-ubsan
> Rev af4cfb1e2cdae70cbda8ff73eca8724cb4538dd6 / Sun, 24 Mar 2019 14:52:01
> +0000
> Lib [2019-03-25 10:04:36 UTC]: fail
> See
> boost/bin.v2/libs/beast/clang-darwin-ubsan/debug/cxxstd-11-iso/link-static/threading-multi/visibility-hidden

ok, this says it fails. Fine.

Clicking on that link gives you the following output:

Test output:
> boost/bin.v2/libs/beast/clang-darwin-ubsan/debug/cxxstd-11-iso/link-static/threading-multi/visibility-hidden
> Rev af4cfb1e2cdae70cbda8ff73eca8724cb4538dd6 / Sun, 24 Mar 2019 14:52:01
> +0000
> Link [2019-03-25 10:04:36 UTC]: succeed
>
> "ar" rc
> "/Sources/boost/testing/results/boost/bin.v2/libs/beast/clang-darwin-ubsan/debug/cxxstd-11-iso/link-static/threading-multi/visibility-hidden/libstatic_beast.a"
> "/Sources/boost/testing/results/boost/bin.v2/libs/beast/clang-darwin-ubsan/debug/cxxstd-11-iso/link-static/threading-multi/visibility-hidden/test/_lib_beast.o"
> "ranlib" "-cs"
> "/Sources/boost/testing/results/boost/bin.v2/libs/beast/clang-darwin-ubsan/debug/cxxstd-11-iso/link-static/threading-multi/visibility-hidden/libstatic_beast.a"
>

That says that it succeeded.
However, the previous page shows that it failed.
Which is it? I don't know.

Any suggestion on how to figure this out / improve the results?

-- Marshall



Boost-testing list run by mbergal at meta-comm.com