possible ruby fix

[Thomas Klausner <wiz%netbsd.org@localhost>]

Index: Makefile
===================================================================
RCS file: /cvsroot/pkgsrc/lang/ruby40/Makefile,v
retrieving revision 1.2
diff -u -r1.2 Makefile
--- Makefile	5 May 2026 14:09:07 -0000	1.2
+++ Makefile	11 May 2026 07:39:29 -0000
@@ -11,7 +11,7 @@
 
 RUBY_VERSIONS_ACCEPTED=	40
 
-MAKE_JOBS_SAFE=		no
+#MAKE_JOBS_SAFE=		no
 USE_GCC_RUNTIME=	yes
 USE_LANGUAGES=		c c++
 USE_TOOLS+=		pax yacc pkg-config
Index: distinfo
===================================================================
RCS file: /cvsroot/pkgsrc/lang/ruby40/distinfo,v
retrieving revision 1.5
diff -u -r1.5 distinfo
--- distinfo	5 May 2026 14:09:07 -0000	1.5
+++ distinfo	11 May 2026 07:39:29 -0000
@@ -17,6 +17,7 @@
 SHA1 (patch-lib_rubygems_installer.rb) = 68a43121b4faab060afa4e352e1603f9d6630a47
 SHA1 (patch-lib_rubygems_platform.rb) = 6bd2e1c5c3064f9d694508169a5bf0a594d7bbef
 SHA1 (patch-lib_rubygems_specification.rb) = f543cab6b56fec0f7195d4eabddedc080c18688c
+SHA1 (patch-template_Makefile.in) = 1159c28d70a5e8ab8e4b1aabb7059c8f2e4e4839
 SHA1 (patch-test_rubygems_test__gem.rb) = 32f7c7d7f8a024c045d78c2bce93944fc3113d04
 SHA1 (patch-thread__pthread.c) = 7c1231933a2d6ce9d56891ab512371841697fbca
 SHA1 (patch-tool_ifchange) = 1803bb6a1836e232dcabcf38f11c9881dbf726ea
Index: patches/patch-template_Makefile.in
===================================================================
RCS file: patches/patch-template_Makefile.in
diff -N patches/patch-template_Makefile.in
--- /dev/null	1 Jan 1970 00:00:00 -0000
+++ patches/patch-template_Makefile.in	11 May 2026 07:39:29 -0000
@@ -0,0 +1,28 @@
+$NetBSD$
+
+Try fixing make parallelism/submake problems.
+
+1. make sure no stale probes.o is there when the probes.o is rebuilt
+2. ruby-glommed and probes.stamp depend on the same files, make sure
+   they don't trigger the same objects twice.
+
+--- template/Makefile.in.orig	2026-05-07 20:25:14.644318083 +0000
++++ template/Makefile.in
+@@ -523,7 +523,7 @@ probes.stamp: $(DTRACE_REBUILD_OBJS)
+ 
+ probes.stamp: $(DTRACE_REBUILD_OBJS)
+ 	$(Q) if test -f $@ -o -f probes.$(OBJEXT); then \
+-	  $(RM) $(DTRACE_REBUILD_OBJS) $@; \
++	  $(RM) $(DTRACE_REBUILD_OBJS) $@ probes.$(OBJEXT); \
+ 	  $(ECHO0) "rebuilding objects which were modified by \"dtrace -G\""; \
+ 	  $(MAKE) $(DTRACE_REBUILD_OBJS); \
+ 	fi
+@@ -537,7 +537,7 @@ probes.$(OBJEXT): $(srcdir)/probes.d $(DTRACE_REBUILD:
+ 
+ # DTrace static library hacks described here:
+ # https://marc.info/?l=opensolaris-dtrace-discuss&m=114761203110734&w=4
+-ruby-glommed.$(OBJEXT):
++ruby-glommed.$(OBJEXT): probes.$(OBJEXT)
+ 	@$(ECHO) generating a glommed object with DTrace probes for static library
+ 	$(Q) $(LD) -r -o $@ $(OBJS)
+ 

Hi!

I tried looking at the build failures like this one:

http://www.ki.nu/pkgsrc/reports/current/NetBSD-9.0/20260506.1500/ruby40-4.0.3/install.log

i.e.

processing probes in object files
dtrace: target object (probes.o) already exists. Please remove the target
dtrace: object and rebuild all the source objects if you wish to run the DTrace
dtrace: linking process again
linking miniruby
ld: array.o: in function `empty_ary_alloc':
array.c:(.text+0x11b): undefined reference to `__dtraceenabled_ruby___array__create'
ld: array.c:(.text+0x17a): undefined reference to `__dtrace_ruby___array__create'
ld: array.o: in function `ary_new':
...

I'm running the attached patch locally and haven't seen the problem
again, but would like some more testing before committing it first. If
you're affected, can you please try it?

Thanks,
 Thomas