[clang][Driver] Add flag for setting SkipFunctionBodies

[kadircet]

This is an option set by certain tools (clangd and ASTUnit). Sometimes
there are crashes in clang unique to this configuration and it's really
hard to provide reproducers without invoking the tool.

[llvmbot]

@llvm/pr-subscribers-clang

Author: kadir çetinkaya (kadircet)

Changes

This is an option set by certain tools (clangd and ASTUnit). Sometimes
there are crashes in clang unique to this configuration and it's really
hard to provide reproducers without invoking the tool.

---

Full diff: https://github.com/llvm/llvm-project/pull/100135.diff

2 Files Affected:

• (modified) clang/include/clang/Driver/Options.td (+3)
• (added) clang/test/Frontend/skip-function-bodies.cc (+8)

diff --git a/clang/include/clang/Driver/Options.td b/clang/include/clang/Driver/Options.td
index 69269cf7537b0..130b2bd68c8b4 100644
--- a/clang/include/clang/Driver/Options.td
+++ b/clang/include/clang/Driver/Options.td
@@ -7525,6 +7525,9 @@ def code_completion_brief_comments : Flag<["-"], "code-completion-brief-comments
 def code_completion_with_fixits : Flag<["-"], "code-completion-with-fixits">,
   HelpText<"Include code completion results which require small fix-its.">,
   MarshallingInfoFlag<FrontendOpts<"CodeCompleteOpts.IncludeFixIts">>;
+def skip_function_bodies : Flag<["-"], "skip-function-bodies">,
+  HelpText<"Skip function bodies when possible">,
+  MarshallingInfoFlag<FrontendOpts<"SkipFunctionBodies">>;
 def disable_free : Flag<["-"], "disable-free">,
   HelpText<"Disable freeing of memory on exit">,
   MarshallingInfoFlag<FrontendOpts<"DisableFree">>;
diff --git a/clang/test/Frontend/skip-function-bodies.cc b/clang/test/Frontend/skip-function-bodies.cc
new file mode 100644
index 0000000000000..dd5ad21424bac
--- /dev/null
+++ b/clang/test/Frontend/skip-function-bodies.cc
@@ -0,0 +1,8 @@
+// Trivial check to ensure skip-function-bodies flag is propagated.
+//
+// RUN: %clang_cc1 -verify -skip-function-bodies -pedantic-errors %s
+
+int f() {
+  // normally this should emit some diags, but we're skipping it!
+  this is garbage; 
+}

[ilya-biryukov]

I am generally supportive of the idea to expose this for more easily reproducing bugs in source tools, and clangd, in particular.
However, I think it'd be useful to make sure the Clang maintainers are on board.
@cor3ntin, @AaronBallman do you have any objection to exposing this option?

[ilya-biryukov]

Let's give other folks a chance to reply before landing this.

[AaronBallman]

My initial thinking was that this doesn't scale particularly well as there may be other options we want to also set at the same time to help produce a reproducer. But after looking through the existing options, I don't really see others that don't already have some form of option. But I'll still ask the question: would it make sense to name this -helpful-reproducers or something more generic so we don't feel tempted to expose even more internal options this way?

That said, so long as this is a -cc1-only option, it seems reasonable to me.

[MaskRay]

The subject could be clarified: "Add cc1 option for ..." so that readers are clear this is not a driver option.

[kadircet]

That said, so long as this is a -cc1-only option, it seems reasonable to m

this is a cc1-only option. added test, updated the commit message.

[AaronBallman]

LGTM

[AaronBallman]

Yay for a diagnostic, boo for it suggesting -Xclang in this instance, but I think we can live with it.

[ilya-biryukov]

But I'll still ask the question: would it make sense to name this -helpful-reproducers or something more generic so we don't feel tempted to expose even more internal options this way?

This is tricky as some issues reproduce only when we skip function bodies and sometimes when we don't. Sometimes we need a PCM that skips function bodies and another file consuming it that does not (this is a normal operation mode of Clangd).
So even this particular flag does not bundle well with any other as we might need to set it independently of everything else.

+1 to Fangrui's suggestion to mention this is a -cc1 flag in the subject line when committing this.
Otherwise, LGTM.