[libcxx-commits] [PATCH] D84801: [libcxx] [test] XFAIL midpoint.float.pass on AArch64 with compiler-rt

[Sergej Jaskiewicz via Phabricator via libcxx-commits]

broadwaylamb created this revision.
broadwaylamb added a reviewer: ldionne.
Herald added subscribers: libcxx-commits, danielkiss, dexonsmith, kristof.beyls, dberris.
Herald added a project: libc++.
Herald added a reviewer: libc++.
broadwaylamb requested review of this revision.

This test makes use of subnormal long doubles, which are not implemented in compiler-rt, at least for the aarch64-linux-gnu triple.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D84801

Files:
  libcxx/test/lit.site.cfg.in
  libcxx/test/std/numerics/numeric.ops/numeric.ops.midpoint/midpoint.float.pass.cpp
  libcxx/utils/libcxx/test/config.py


Index: libcxx/utils/libcxx/test/config.py
===================================================================
--- libcxx/utils/libcxx/test/config.py
+++ libcxx/utils/libcxx/test/config.py
@@ -276,6 +276,12 @@
         if libcxx_gdb and 'NOTFOUND' not in libcxx_gdb:
             self.config.available_features.add('libcxx_gdb')
             self.cxx.libcxx_gdb = libcxx_gdb
+        
+        if self.config.target_triple:
+            self.config.available_features.add(self.config.target_triple)
+
+        if self.get_lit_conf('use_compiler_rt'):
+            self.config.available_features.add('use_compiler_rt')
 
     def configure_compile_flags(self):
         self.configure_default_compile_flags()
Index: libcxx/test/std/numerics/numeric.ops/numeric.ops.midpoint/midpoint.float.pass.cpp
===================================================================
--- libcxx/test/std/numerics/numeric.ops/numeric.ops.midpoint/midpoint.float.pass.cpp
+++ libcxx/test/std/numerics/numeric.ops/numeric.ops.midpoint/midpoint.float.pass.cpp
@@ -7,6 +7,10 @@
 //===----------------------------------------------------------------------===//
 //
 // UNSUPPORTED: c++03, c++11, c++14, c++17
+//
+// subnormal floats are not implemented in compiler-rt builtins for AArch64.
+// XFAIL: aarch64-linux-gnu && use_compiler_rt
+//
 // <numeric>
 
 // template <class _Float>
Index: libcxx/test/lit.site.cfg.in
===================================================================
--- libcxx/test/lit.site.cfg.in
+++ libcxx/test/lit.site.cfg.in
@@ -32,6 +32,7 @@
 config.executor                 = "@LIBCXX_EXECUTOR@"
 config.llvm_unwinder            = @LIBCXXABI_USE_LLVM_UNWINDER@
 config.builtins_library         = "@LIBCXX_BUILTINS_LIBRARY@"
+config.use_compiler_rt          = "@LIBCXX_USE_COMPILER_RT@"
 config.has_libatomic            = @LIBCXX_HAS_ATOMIC_LIB@
 config.debug_build              = @LIBCXX_DEBUG_BUILD@
 config.libcxxabi_shared         = @LIBCXX_LINK_TESTS_WITH_SHARED_LIBCXXABI@


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D84801.281363.patch
Type: text/x-patch
Size: 1983 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/libcxx-commits/attachments/20200728/f7f8e6c2/attachment-0001.bin>