[PR52549][clang-cl] Predefine _MSVC_EXECUTION_CHARACTER_SET

[zero9178]

Summary

Since VS 2022 17.1 MSVC predefines _MSVC_EXECUTION_CHARACTER_SET to inform the users of the execution character set defined at compile time. The value the macro expands to is a Windows Code Page Identifier which are documented here: https://docs.microsoft.com/en-us/windows/win32/intl/code-page-identifiers

As clang currently only supports UTF-8 it is defined as 65001. If clang-cl were to support a different execution character set in the future we'd have to change the value.

Fixes https://bugs.llvm.org/show_bug.cgi?id=52549

[mstorsjo]

Would it be possible to move setting of this flag to somewhere else (e.g. somewhere in Driver/ToolChains/MSVC.cpp), so that it would be picked up also when building with the gcc-style driver with --target=*-windows-msvc? Overall in earlier discussions, there's been mostly positive reception to making that mode more usable even if it isn't quite as supported as actually invoking clang-cl. (For some things, like the equivalent of the cl level -MD/-MT options, there's currently nothing available in the gcc style driver - but for a feature like this, I don't see any reason why it shouldn't be available everywhere.)

[zero9178]

Enable the macro for all MSVC targets, regardless of driver in use.

[mstorsjo]

Thanks! This LGTM - but maybe wait a while if someone else has something to say. (Although I guess americans are on holiday today, so there's probably not many that will see and care today.)

[CaseyCarter]

LGTM - thanks for the quick response!

[zero9178]

Seems like this is causing failures in a test for clang-tidy if the triple used is an MSVC one. Let me see what I can do about it

[zero9178]

Alternative implementation of the original patch.

Moves the defining macro from being added to the cc1 command line by the driver, to being a builtin macro. This should avoid the issues with clang-tidy.

[rnk]

In D114576#3153189, @mstorsjo wrote:

Would it be possible to move setting of this flag to somewhere else (e.g. somewhere in Driver/ToolChains/MSVC.cpp), so that it would be picked up also when building with the gcc-style driver with --target=*-windows-msvc? Overall in earlier discussions, there's been mostly positive reception to making that mode more usable even if it isn't quite as supported as actually invoking clang-cl. (For some things, like the equivalent of the cl level -MD/-MT options, there's currently nothing available in the gcc style driver - but for a feature like this, I don't see any reason why it shouldn't be available everywhere.)

+1, that was going to be my suggestion.

clang/lib/Driver/ToolChains/Clang.cpp

if (Args.hasArg(options::OPT__SLASH_Zl)) {
     CmdArgs.push_back(FlagForCRT.data());
 
     // This provides POSIX compatibility (maps 'open' to '_open'), which most
     // users want.  The /Za flag to cl.exe turns this off, but it's not
     // implemented in clang.
     CmdArgs.push_back("--dependent-lib=oldnames");
   }
 
+  // Starting with VS 2022 17.1, MSVC predefines the below macro to inform
+  // users of the execution character set defined at compile time.
+  // The value given is the Windows Code Page Identifier:
+  // https://docs.microsoft.com/en-us/windows/win32/intl/code-page-identifiers
+  //
+  // Clang currently only supports UTF-8, so we'll use 65001
+  CmdArgs.push_back("-D_MSVC_EXECUTION_CHARACTER_SET=65001");
+
   if (Arg *ShowIncludes =
           Args.getLastArg(options::OPT__SLASH_showIncludes,
                           options::OPT__SLASH_showIncludes_user)) {
     CmdArgs.push_back("--show-includes");
     if (ShowIncludes->getOption().matches(options::OPT__SLASH_showIncludes))
       CmdArgs.push_back("-sys-header-deps");
   }
 

clang/test/Driver/cl-execution-character-set.c

+// Note: %s must be preceded by --, otherwise it may be interpreted as a
+// command-line option, e.g. on Mac where %s is commonly under /Users.
+
+// Only UTF-8 is currently supported, which has the Windows Code
+// Page Identifier 65001
+// RUN: %clang_cl -### -- %s 2>&1 | FileCheck %s
+// CHECK: -D_MSVC_EXECUTION_CHARACTER_SET=65001