plotfi created this revision.
plotfi added reviewers: rjmccall, ab.
Herald added a project: clang.
Herald added a subscriber: cfe-commits.
After speaking with @ab and @rjmccall on
https://reviews.llvm.org/D72841#2027740 and
https://github.com/apple/llvm-project/pull/1202.
it sounds like code-gen option parsing should depend on language-option
parsing, not the other way around. Apple Clang llvm-project master does this in
the right order but upstream llvm.org does it in the opposite order.
This patch changes the order to the way apple-master does it, which we think is
the correct order. At the moment D72841 <https://reviews.llvm.org/D72841>
causes some bot failures on apple-master (only on tests added in D72841
<https://reviews.llvm.org/D72841>). I think D72841
<https://reviews.llvm.org/D72841> should be reverted, and then reapplied after
_this_ patch is landed.
Repository:
rG LLVM Github Monorepo
https://reviews.llvm.org/D79730
Files:
clang/lib/Frontend/CompilerInvocation.cpp
Index: clang/lib/Frontend/CompilerInvocation.cpp
===================================================================
--- clang/lib/Frontend/CompilerInvocation.cpp
+++ clang/lib/Frontend/CompilerInvocation.cpp
@@ -3626,8 +3626,6 @@
InputKind DashX = ParseFrontendArgs(Res.getFrontendOpts(), Args, Diags,
LangOpts.IsHeaderFile);
ParseTargetArgs(Res.getTargetOpts(), Args, Diags);
- Success &= ParseCodeGenArgs(Res.getCodeGenOpts(), Args, DashX, Diags,
- Res.getTargetOpts(), Res.getFrontendOpts());
ParseHeaderSearchArgs(Res.getHeaderSearchOpts(), Args,
Res.getFileSystemOpts().WorkingDir);
llvm::Triple T(Res.getTargetOpts().Triple);
@@ -3665,6 +3663,9 @@
LangOpts.FunctionAlignment =
getLastArgIntValue(Args, OPT_function_alignment, 0, Diags);
+ Success &= ParseCodeGenArgs(Res.getCodeGenOpts(), Args, DashX, Diags,
+ Res.getTargetOpts(), Res.getFrontendOpts());
+
if (LangOpts.CUDA) {
// During CUDA device-side compilation, the aux triple is the
// triple used for host compilation.
Index: clang/lib/Frontend/CompilerInvocation.cpp
===================================================================
--- clang/lib/Frontend/CompilerInvocation.cpp
+++ clang/lib/Frontend/CompilerInvocation.cpp
@@ -3626,8 +3626,6 @@
InputKind DashX = ParseFrontendArgs(Res.getFrontendOpts(), Args, Diags,
LangOpts.IsHeaderFile);
ParseTargetArgs(Res.getTargetOpts(), Args, Diags);
- Success &= ParseCodeGenArgs(Res.getCodeGenOpts(), Args, DashX, Diags,
- Res.getTargetOpts(), Res.getFrontendOpts());
ParseHeaderSearchArgs(Res.getHeaderSearchOpts(), Args,
Res.getFileSystemOpts().WorkingDir);
llvm::Triple T(Res.getTargetOpts().Triple);
@@ -3665,6 +3663,9 @@
LangOpts.FunctionAlignment =
getLastArgIntValue(Args, OPT_function_alignment, 0, Diags);
+ Success &= ParseCodeGenArgs(Res.getCodeGenOpts(), Args, DashX, Diags,
+ Res.getTargetOpts(), Res.getFrontendOpts());
+
if (LangOpts.CUDA) {
// During CUDA device-side compilation, the aux triple is the
// triple used for host compilation.
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
