Date: Mon, 21 Feb 2005 23:36:34 -0600 From: "Jeremy Messenger" <mezz7@cox.net> To: "Garance A Drosihn" <drosih@rpi.edu> Cc: ports@freebsd.org Subject: Re: FreeBSD Port: ruby18-1.8.2_2 Message-ID: <opsmk268vc9aq2h7@mezz.mezzweb.com> In-Reply-To: <p0621020abe40281a2ae3@[128.113.24.47]> References: <64348.207.219.213.162.1109011818.squirrel@207.219.213.162> <a5d5f155be4e65d41f9775df96135160@khera.org> <20050221194954.GA49438@xor.obsecurity.org> <opsmkkzit99aq2h7@mezz.mezzweb.com> <6a2d0a658b1cfe9503183eb978e4e2d6@snsonline.net> <opsmklbltj9aq2h7@mezz.mezzweb.com> <p0621020abe40281a2ae3@[128.113.24.47]>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, 21 Feb 2005 20:24:06 -0500, Garance A Drosihn <drosih@rpi.edu> wrote: > At 5:10 PM -0600 2/21/05, Jeremy Messenger wrote: >> >> My lastest patch was: http://people.freebsd.org/~mezz/diff/ruby18.diff >> > > I tried this on powerPC, and ruby built fine. I haven't tested > it much, but the things I tested seem to work fine. > > I tried it on sparc64, and ruby didn't build. The attempt > ended with: <snip> Ruby has few '__sparc__' and '__sparc_v9__' in the very few places, but none of '__sparc64__'. Are there any chance that it will need to add a new define, '__sparc64__', to make ruby works better? I am lacking on knowlege of programming, so someone with the programming knowledge should take a look at those. ======================================= # find . -name \* | xargs egrep -l "sparc" ./work/ruby-1.8.2/config.guess ./work/ruby-1.8.2/config.sub ./work/ruby-1.8.2/configure.in ./work/ruby-1.8.2/defines.h ./work/ruby-1.8.2/dln.c ./work/ruby-1.8.2/gc.c ./work/ruby-1.8.2/doc/ChangeLog-1.8.0 ./work/ruby-1.8.2/ext/digest/sha1/sha1.c ./work/ruby-1.8.2/configure ./work/ruby-1.8.2/configure.lineno ./work/ruby-1.8.2/configure.bak ======================================= Looks like about four to five files that need to take a look at. Cheers, Mezz > Looking in the work directory, it seems the bus error came up > while running mini-ruby: > > work/ruby-1.8.2/miniruby.core > > gdb tells me: > > Core was generated by `miniruby'. > Program terminated with signal 10, Bus error. > > (gdb) backtrace full > #0 0x0000000040859d5c in __vfprintf () from /lib/libc.so.6 > > I noticed the other email message which talked about the settings > of CFLAGS, so I changed my /etc/make.conf > from: > CFLAGS= -O -pipe > to: > CFLAGS+= -O -pipe -g > > This changed the resulting 'cc' commands > from: > cc -O -pipe -fPIC -I. -I. -c main.c > to: > cc -O2 -fno-strict-aliasing -pipe -O -pipe -g -fPIC -I. -I. -c > main.c > > This still core-dumps. I wondered if it might be significant that > "something" wanted this to be -O2, but my CFLAGS was setting a -O. > So I moved the line for '.include <bsd.port.pre.mk>' such that it > was before any of the lines which set CFLAGS in the makefile (I > put it right after GNU_CONFIGURE=yes). That didn't change anything. > All the 'cc' commands came out the same, and mini-ruby core-dumped. > > So then I tried commenting-out the setting of CFLAGS in my > /etc/make.conf file. That gave me 'cc' commands of: > > cc -O2 -fno-strict-aliasing -pipe -fPIC -I. -I. -c main.c > > but it still just core-dumps. I then tried to make it with DEBUG=yes, > and used gdb to step through it. It does start in main, and it > executes several thousand instructions before it dies. -- mezz7@cox.net - mezz@FreeBSD.org FreeBSD GNOME Team http://www.FreeBSD.org/gnome/ - gnome@FreeBSD.org
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?opsmk268vc9aq2h7>