Hi, strange, I think the __sync_fetch_and_add_4 comes from defaulting to i386 or so and should go away when you specify some -mcpu=i486 (or higher). Maybe when not build inside T2 gcc picks the current host as default. I have to check.
Greetings, René On 10.12.2009, at 03:40, Barry Kauler wrote: > Hi, > There's only one error, and we only get it with (tying to) compile > certain packages. It is "__sync_fetch_and_add_4 undefined". Googling > around, that is a common error message, it seems typically in > cross-compile situations, but I didn't find any solutions, except to > recompile gcc. > > It's a problem that we have always had. Back in 2007 when I used T2 to > compile the base packages for the Puppy 4.x series, I had to do the > same thing. > Another Puppy developer, Kirk, uses T2 and he was the one who put me > onto the solution, that is, to recompile gcc outside of T2. > > Regards, > Barry Kauler > > > > On 12/9/09, René Rebe <[email protected]> wrote: >> Hi Barry, >> >> Barry Kauler wrote: >>> Rene and anyone else interested, >>> >>> I have recently been through the exercise of compiling with T2, to get >>> binary packages that can be used for the next Puppy Linux derivative, >>> that I have named Quirky. >>> >>> I have created a page on T2 that explains how it is all done: >>> >>> http://bkhome.org/t2/ >>> >>> My Woof distro build system is able to import binary packages created >>> by T2 and post-process them into cut-down packages, plus a package >>> database file is created in the Puppy Standard Format. The Puppy >>> Package Manager is also able to install T2 .tar.bz2 binary packages. >>> Woof is introduced here: >>> >>> http://bkhome.org/woof/ >>> >>> I have released Quirky version 0.0.1. >>> Release announcement: >>> http://distro.ibiblio.org/pub/linux/distributions/quirky/quirky-001/001-readme.htm >>> Download ISO: >>> http://distro.ibiblio.org/pub/linux/distributions/quirky/quirky-001/ >>> Home page: http://bkhome.org/quirky/ >>> >> >> Nice! And thanks for letting us and your users know! >> >>> One thing to note. A problem that we have always had with T2 is that >>> gcc is slightly broken when used outside the T2 environment. I have >>> always had to recompile it, as I have documented on the T2 page. I >>> don't know why... I thought it might have something to do with the >>> configure options in T2 specifying all three '--build', '--host' and >>> '--target', whereas I just specify '--build'. T2 configure: >>> >>> ../configure --prefix=/usr --bindir=/usr/bin --sbindir=/usr/sbin >>> --libdir=/usr/lib --datadir=/usr/share --includedir=/usr/include >>> --infodir=/usr/info --mandir=/usr/man --sysconfdir=/etc >>> --localstatedir=/var --disable-debug --without-libpam --without-pam >>> --disable-libpam --disable-pam --build=i486-t2-linux-gnu >>> --host=i486-t2-linux-gnu --target=i486-t2-linux-gnu >>> --enable-__cxa_atexit --disable-checking --disable-bootstrap >>> --disable-libstdcxx-pch --disable-multilib --with-gnu-as --with-gnu-ld >>> --enable-threads=posix --disable-libgcj >>> --enable-languages=c,c++,objc,fortran --enable-shared >>> >>> My configure, compiled in a running Quirky: >>> >>> ../configure --prefix=/usr --bindir=/usr/bin --sbindir=/usr/sbin >>> --libdir=/usr/lib --datadir=/usr/share --includedir=/usr/include >>> --infodir=/usr/info --mandir=/usr/man --sysconfdir=/etc >>> --localstatedir=/var --disable-debug --without-libpam --without-pam >>> --disable-libpam --disable-pam --build=i486-t2-linux-gnu >>> --enable-__cxa_atexit --disable-checking --disable-bootstrap >>> --disable-libstdcxx-pch --disable-multilib --with-gnu-as --with-gnu-ld >>> --enable-threads=posix --disable-libgcj >>> --enable-languages=c,c++,fortran --enable-shared >>> >> Strange, I never had a problem with our GCC when running it without the >> T2 environment from the command line (e.g. to build my or other sources >> manually). What kind of errors / issue do you experience with the T2 >> built GCC? >> >> Best regards, >> René >> >> -- >> René Rebe, ExactCODE GmbH, Jaegerstr. 67, DE-10785 Berlin >> http://exactcode.de | http://t2-project.org | http://rene.rebe.name >> >> -- René Rebe, ExactCODE GmbH, Jaegerstr. 67, DE-10117 Berlin http://exactcode.com | http://t2-project.org | http://rene.rebe.name
----------------------------------------------------------- If you wish to unsubscribe from this mailing, send mail to [email protected] with a subject of: unsubscribe t2
