I'm trying to tell the Boost regression matrix that GCC 8 failures are to be expected using explicit-failures-markup.xml, but it seems to not be working: https://www.boost.org/development/tests/master/developer/outcome.html Here is my explicit-failures-markup.xml: <?xml version="1.0" encoding="utf-8"?> <explicit-failures-markup> <!-- outcome --> <library name="outcome"> <mark-unusable> <toolset name="clang-linux-3*"/> <toolset name="gcc-3*"/> <toolset name="gcc-4*"/> <toolset name="gcc-5*"/> <toolset name="gcc-mngw-3*"/> <toolset name="gcc-mngw-4*"/> <toolset name="gcc-mngw-5*"/> <toolset name="msvc-7*"/> <toolset name="msvc-8*"/> <toolset name="msvc-9*"/> <toolset name="msvc-10*"/> <toolset name="msvc-11*"/> <toolset name="msvc-12*"/> <toolset name="msvc-13*"/> <toolset name="msvc-14.0.*"/> <note author="Niall Douglas">Complete C++14 support is the minimum requirement.</note> </mark-unusable> <mark-expected-failures> <toolset name="gcc-8*"/> <note author="Niall Douglas">GCC 8 currently ICEs with Outcome's unit tests in C++ 17 or later.</note> </mark-expected-failures> </library> </explicit-failures-markup> What am I doing wrong? Thanks, Niall
Niall Douglas wrote:
I'm trying to tell the Boost regression matrix that GCC 8 failures are to be expected using explicit-failures-markup.xml, but it seems to not be working:
https://www.boost.org/development/tests/master/developer/outcome.html
...
What am I doing wrong?
When you update your local explicit-failures-markup.xml, a cron script is supposed to pick that up and update status/explicit-failures-markup.xml, as in https://github.com/boostorg/boost/commit/2acdfb0253a0a9bbab04d700364f330dbd9.... Unfortunately, this doesn't seem to occur in your case for some reason. I think that the person to contact for this is Daniel James.
When you update your local explicit-failures-markup.xml, a cron script is supposed to pick that up and update status/explicit-failures-markup.xml, as in https://github.com/boostorg/boost/commit/2acdfb0253a0a9bbab04d700364f330dbd9....
Unfortunately, this doesn't seem to occur in your case for some reason. I think that the person to contact for this is Daniel James.
Excellent. I had been beginning to doubt myself as an engineer. I'll drop him a line. Niall
participants (2)
-
Niall Douglas
-
Peter Dimov