> On Oct 11, 2018, at 3:30 PM, Stella Stamenova via Phabricator > <revi...@reviews.llvm.org> wrote: > > stella.stamenova added a comment. > > I spent a bit more time on this and I think I have a better idea of what's > happening. > > Vedant added functionality so that we can use FileCheck-style checks inside > non-lit LLDB tests. Part of that change was to require a parameter > --filecheck to be passed when calling the scripts (such as dotests.py). Also > as part of that change, the CMakeFiles now propagate a default location for > FileCheck as an input to the scripts, so that if you just run the tests by > default using any of the generated outputs (for ninja, or make, or VS), they > work correctly > > The bots (at least some bots) don't just use the generated solutions to run > the tests. There's some complicated logic in the zorg repo to create the list > of parameters that are passed to the bots and they do not include the new > --filecheck argument.
I did update zorg. See r342507 and r342559. Ah, is there a Windows-specific configuration I missed? > So it's not individual bots that need to be updated, but the test harness. I > haven't really looked at it before, so I'm still trying to understand it - > especially how we would get FileCheck to any of the bots that run the tests > remotely using lldbserver today. zorg/jenkins/build.py in https://github.com/llvm-mirror/zorg.git <https://github.com/llvm-mirror/zorg.git> might be relevant. vedant > > > https://reviews.llvm.org/D53175 > > >
_______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
