#15635 closed defect (fixed)
Ruby 1.8.7 dumps core on 10.5 ppc
Reported by: | blair (Blair Zajac) | Owned by: | pguyot (Paul Guyot) |
---|---|---|---|
Priority: | Normal | Milestone: | |
Component: | ports | Version: | 1.6.0 |
Keywords: | Cc: | febeling@… | |
Port: |
Description
I can compile and test ruby 1.8.7 fine on 10.5.3 i386, but it dumps core on 10.5.3 ppc:
/usr/bin/gcc-4.0 -O2 -fno-common -pipe -fno-common -DRUBY_EXPORT -I. -I. -I/opt/local/include -D_XOPEN_SOURCE -D_DARWIN_C_SOURCE -c string.c /usr/bin/gcc-4.0 -O2 -fno-common -pipe -fno-common -DRUBY_EXPORT -I. -I. -I/opt/local/include -D_XOPEN_SOURCE -D_DARWIN_C_SOURCE -c struct.c /usr/bin/gcc-4.0 -O2 -fno-common -pipe -fno-common -DRUBY_EXPORT -I. -I. -I/opt/local/include -D_XOPEN_SOURCE -D_DARWIN_C_SOURCE -c time.c /usr/bin/gcc-4.0 -O2 -fno-common -pipe -fno-common -DRUBY_EXPORT -I. -I. -I/opt/local/include -D_XOPEN_SOURCE -D_DARWIN_C_SOURCE -c util.c /usr/bin/gcc-4.0 -O2 -fno-common -pipe -fno-common -DRUBY_EXPORT -I. -I. -I/opt/local/include -D_XOPEN_SOURCE -D_DARWIN_C_SOURCE -c variable.c /usr/bin/gcc-4.0 -O2 -fno-common -pipe -fno-common -DRUBY_EXPORT -I. -I. -I/opt/local/include -D_XOPEN_SOURCE -D_DARWIN_C_SOURCE -c version.c /usr/bin/gcc-4.0 -O2 -fno-common -pipe -fno-common -DRUBY_EXPORT -I. -I. -I/opt/local/include -D_XOPEN_SOURCE -D_DARWIN_C_SOURCE -c ./missing/alloca.c ./missing/alloca.c:144: warning: conflicting types for built-in function 'alloca' /usr/bin/gcc-4.0 -O2 -fno-common -pipe -fno-common -DRUBY_EXPORT -I. -I. -I/opt/local/include -D_XOPEN_SOURCE -D_DARWIN_C_SOURCE -c dmyext.c ar rcu libruby-static.a array.o bignum.o class.o compar.o dir.o dln.o enum.o enumerator.o error.o eval.o file.o gc.o hash.o inits.o io.o marshal.o math.o numeric.o object.o pack.o parse.o process.o prec.o random.o range.o re.o regex.o ruby.o signal.o sprintf.o st.o string.o struct.o time.o util.o variable.o version.o alloca.o dmyext.o /usr/bin/gcc-4.0 -O2 -fno-common -pipe -fno-common -DRUBY_EXPORT -I. -I. -I/opt/local/include -D_XOPEN_SOURCE -D_DARWIN_C_SOURCE -c main.c /usr/bin/gcc-4.0 -O2 -fno-common -pipe -fno-common -DRUBY_EXPORT -I. -I. -I/opt/local/include -D_XOPEN_SOURCE -D_DARWIN_C_SOURCE -c dmydln.c /usr/bin/gcc-4.0 -O2 -fno-common -pipe -fno-common -DRUBY_EXPORT -L. -L/opt/local/lib main.o dmydln.o libruby-static.a -lpthread -ldl -lobjc -o miniruby ./ext/purelib.rb:5: [BUG] Segmentation fault ruby 1.8.7 (2008-05-31 patchlevel 0) [powerpc-darwin9.3.0] make: *** [.rbconfig.time] Abort trap (core dumped) Warning: the following items did not execute (for ruby): org.macports.build Error: Status 1 encountered during processing.
Change History (11)
comment:1 Changed 16 years ago by blair (Blair Zajac)
comment:2 Changed 16 years ago by raimue (Rainer Müller)
Did you already report this upstream? I think this is out of range what we can fix easily, so upstream should have a look.
comment:3 Changed 16 years ago by blair (Blair Zajac)
Me, no, I didn't report it upstream.
It's the port maintainer's job to handle the MacPorts issue, doing whatever it takes to have a successful MacPorts build, such as reporting it upstream, finding a patch, etc.
comment:5 Changed 16 years ago by febeling@…
Hm, does still segfault on PPC/10.5.3 after update to p17, unfortunately.
comment:6 Changed 16 years ago by febeling@…
It does build and install without error when you remove the "--enable-pthread" line from the portfile.
comment:7 follow-up: 8 Changed 16 years ago by pguyot (Paul Guyot)
From what I understand, it will break tk support.
comment:8 follow-up: 9 Changed 16 years ago by febeling@…
Replying to pguyot@kallisys.net:
From what I understand, it will break tk support.
Yes, I've seen that note as well. I don't know what is best to do.
Maybe it's better to break only Tk than the entire PPC target, and do some dodgy if arch == powerpc thing.
I've filed a bug with the ruby project for this though, because it happens with a pure tar ball as well:
http://redmine.ruby-lang.org/issues/show/193
I mean there is no working ruby 1.8.7 for that machine right now, it seems. But I think most people want it for rails and such, where tk is quite secondary. I think have never invoked a single Tk widget in the 3 years I use ruby.
comment:9 Changed 16 years ago by kimuraw (kimura wataru)
The following change against ruby's Portfile disables "--enable-pthread" on 10.5 ppc Mac. I think this is not so bad to solve the problem temporally.
Index: lang/ruby/Portfile =================================================================== --- lang/ruby/Portfile (revision 783) +++ lang/ruby/Portfile (working copy) @@ -74,6 +74,12 @@ configure.args-delete --enable-pthread } +platform darwin 9 powerpc { + # pthreads breaks ruby-1.8.7 on 10.5 ppc + # http://trac.macports.org/ticket/15635 + configure.args-delete --enable-pthread +} + variant tk conflicts mactk { configure.args-delete --without-tk configure.args-append --with-tk
comment:10 Changed 16 years ago by pguyot (Paul Guyot)
Resolution: | → fixed |
---|---|
Status: | new → closed |
Committed the workaround in r37694 where I also changed the maintainer to kimuraw@.
Some more information. This is the first command that fails from Makefile:
Running it in a debugger to get a stack trace: