tobhe called, he wants his webkitgtk4 back (on powerpc). The diff
below should fix the webkitgtk4 build on powerpc, powerpc64 and
mips64. It's directly inspired from gkoehler's diff for lang/gambit.
The error seen in the latest bulk:
fatal error: error in backend: failed to perform tail call elimination on a
call site marked musttail
Completely untested but also innocuous for other archs. tobhe could
test on bluhm's machine, or maybe gkoehler could put it in his next
build?
ok?
Index: Makefile
===================================================================
RCS file: /home/cvs/ports/www/webkitgtk4/Makefile,v
diff -u -p -r1.212 Makefile
--- Makefile 15 Apr 2024 11:48:20 -0000 1.212
+++ Makefile 15 Apr 2024 15:48:09 -0000
@@ -16,7 +16,7 @@ V = 2.44.1
DISTNAME = webkitgtk-${V}
PKGNAME = webkitgtk${API:S/.//}-${V}
FULLPKGNAME = ${PKGNAME}
-REVISION = 0
+REVISION = 1
EXTRACT_SUFX = .tar.xz
SUBST_VARS = API
Index: patches/patch-Source_WTF_wtf_Compiler_h
===================================================================
RCS file: patches/patch-Source_WTF_wtf_Compiler_h
diff -N patches/patch-Source_WTF_wtf_Compiler_h
--- /dev/null 1 Jan 1970 00:00:00 -0000
+++ patches/patch-Source_WTF_wtf_Compiler_h 15 Apr 2024 15:47:59 -0000
@@ -0,0 +1,15 @@
+Disable musttail attribute, not supported by clang on powerpc, powerpc64
+and mips64.
+
+Index: Source/WTF/wtf/Compiler.h
+--- Source/WTF/wtf/Compiler.h.orig
++++ Source/WTF/wtf/Compiler.h
+@@ -321,7 +321,7 @@
+ /* MUST_TAIL_CALL */
+
+ #if !defined(MUST_TAIL_CALL) && defined(__cplusplus) &&
defined(__has_cpp_attribute)
+-#if __has_cpp_attribute(clang::musttail)
++#if __has_cpp_attribute(clang::musttail) && !defined(__mips__) &&
!defined(__powerpc__)
+ #define MUST_TAIL_CALL [[clang::musttail]]
+ #endif
+ #endif
--
jca
