https://git.ruby-lang.org/ruby.git/commit/?id=41a024f2b9

From 41a024f2b94cea802f7f6d8a1a9de9878aa20c7b Mon Sep 17 00:00:00 2001
From: Alan Wu <alanwu@r...>
Date: Fri, 17 Jun 2022 18:12:55 -0400
Subject: YJIT: Update note about symbol prefixes [ci skip]

---
 yjit.c | 9 ++-------
 1 file changed, 2 insertions(+), 7 deletions(-)

diff --git a/yjit.c b/yjit.c
index fe9098f0b5..a30e3bad94 100644
--- a/yjit.c
+++ b/yjit.c
@@ -56,13 +56,8 @@ STATIC_ASSERT(pointer_tagging_scheme, USE_FLONUM); https://github.com/ruby/ruby/blob/trunk/yjit.c#L56
 // types in C such as int, long, etc. and use `std::os::raw::c_long` and friends on
 // the Rust side.
 //
-// What's up with the long prefix? The "rb_" part is to appease `make leaked-globals`
-// which runs on upstream CI. The rationale for the check is unclear to Alan as
-// we build with `-fvisibility=hidden` so only explicitly marked functions end
-// up as public symbols in libruby.so. Perhaps the check is for the static
-// libruby and or general namspacing hygiene? Alan admits his bias towards ELF
-// platforms and newer compilers.
-//
+// What's up with the long prefix? Even though we build with `-fvisibility=hidden`
+// we are sometimes a static library where the option doesn't prevent name collision.
 // The "_yjit_" part is for trying to be informative. We might want different
 // suffixes for symbols meant for Rust and symbols meant for broader CRuby.
 
-- 
cgit v1.2.1


--
ML: ruby-changes@q...
Info: http://www.atdot.net/~ko1/quickml/