What does |g++-4.6 -v| show?  Sorry I missed the point on putting 
/usr/local/bin ahead in your path.  Also, I'm surprised you had to run brew 
link --overwrite.  Normally that's not necessary.  Did the install tell you to 
do that?  Did it complete successfully?  I've uninstalled and installed it 
quite a few times without having to do anything beyond what I told you.

Just to make sure we're actually working from roughly the same source, if you 
could run the following commands from the root of B2G:

rm -rf objdir-gecko
rm .userconfig
git pull
./repo sync
./config.sh inari
./build.sh clean

That should get you back to an initial state.  If you could also provide the 
output of the following command run from the root of your b2g tree:

cd build && git diff && cd ..

Thanks for sticking with this.

- James


On Wednesday, October 16, 2013 2:45:39 PM UTC-4, Christopher De Cairos wrote:
> I got a bit farther, but I'm hitting this now:
> 
> 
> 
> configure:10426: checking for host gcc c++0x headers bug without rtti
> 
> configure:10438: c++ -c   -std=gnu++0x   conftest.C 1>&5
> 
> cc1plus: error: unrecognized command line option "-std=gnu++0x"
> 
> configure: failed program was:
> 
> #line 10431 "configure"
> 
> #include "confdefs.h"
> 
> #include <memory>
> 
> int main() {
> 
> 
> 
> ; return 0; }
> 
> configure: error: Your host toolchain does not support C++0x/C++11 mode 
> 
> properly. Please upgrade your toolchain
> 
> *** Fix above errors and then restart with               "make -f 
> 
> client.mk build"
> 
> make[3]: *** [configure] Error 1
> 
> make[2]: *** [/Volumes/firefoxos/B2G/objdir-gecko/Makefile] Error 2
> 
> make[1]: *** [build] Error 2
> 
> make: *** [out/target/product/inari/obj/DATA/gecko_intermediates/gecko] 
> 
> Error 2
> 
> 
> 
> Regards,
> 
> 
> 
> Christopher De Cairos
> 
> Integration Engineer - Webmaker
> 
> Mozilla Foundation
> 
> 
> 
> On 2013-10-16 2:21 PM, James Kruth wrote:
> 
> > Christopher,
> 
> >
> 
> > No worries.  Can you verify that the updated brew formula actually 
> > installed /usr/local/bin/g++-4.6?  Also, is /usr/local/bin on your path?
> 
> >
> 
> > - James
> 
> >
> 
> > On Wednesday, October 16, 2013 2:14:47 PM UTC-4, Christopher De Cairos 
> > wrote:
> 
> >> Sorry for the wait, busy day.
> 
> >>
> 
> >>
> 
> >>
> 
> >> With a clean build and following your instructions, I see the same clang
> 
> >>
> 
> >> errors that I was seeing originally.
> 
> >>
> 
> >>
> 
> >>
> 
> >> clang: error: argument unused during compilation: '-I
> 
> >>
> 
> >> frameworks/base/tools/obbtool'
> 
> >>
> 
> >> clang: error: argument unused during compilation: '-I
> 
> >>
> 
> >> out/host/darwin-x86/obj/EXECUTABLES/obbtool_intermediates'
> 
> >>
> 
> >> clang: error: argument unused during compilation: '-I
> 
> >>
> 
> >> dalvik/libnativehelper/include/nativehelper'
> 
> >>
> 
> >> clang: error: argument unused during compilation: '-isystem
> 
> >>
> 
> >> system/core/include'
> 
> >>
> 
> >> clang: error: argument unused during compilation: '-isystem
> 
> >>
> 
> >> hardware/libhardware/include'
> 
> >>
> 
> >> clang: error: argument unused during compilation: '-isystem
> 
> >>
> 
> >> hardware/libhardware_legacy/include'
> 
> >>
> 
> >> clang: error: argument unused during compilation: '-isystem
> 
> >>
> 
> >> hardware/ril/include'
> 
> >>
> 
> >> clang: error: argument unused during compilation: '-isystem
> 
> >>
> 
> >> dalvik/libnativehelper/include'
> 
> >>
> 
> >> clang: error: argument unused during compilation: '-isystem
> 
> >>
> 
> >> frameworks/base/include'
> 
> >>
> 
> >> clang: error: argument unused during compilation: '-isystem
> 
> >>
> 
> >> frameworks/base/opengl/include'
> 
> >>
> 
> >> clang: error: argument unused during compilation: '-isystem
> 
> >>
> 
> >> frameworks/base/native/include'
> 
> >>
> 
> >> clang: error: argument unused during compilation: '-isystem
> 
> >>
> 
> >> external/skia/include'
> 
> >>
> 
> >> clang: error: argument unused during compilation: '-isystem tools/include'
> 
> >>
> 
> >> clang: error: argument unused during compilation: '-isystem
> 
> >>
> 
> >> out/host/darwin-x86/obj/include'
> 
> >>
> 
> >> clang: error: argument unused during compilation: '-include
> 
> >>
> 
> >> system/core/include/arch/darwin-x86/AndroidConfig.h'
> 
> >>
> 
> >> clang: error: argument unused during compilation: '-U DEBUG'
> 
> >>
> 
> >> clang: error: argument unused during compilation: '-U DEBUG'
> 
> >>
> 
> >> clang: error: argument unused during compilation: '-MF
> 
> >>
> 
> >> out/host/darwin-x86/obj/EXECUTABLES/obbtool_intermediates/Main.d'
> 
> >>
> 
> >> make: ***
> 
> >>
> 
> >> [out/host/darwin-x86/obj/EXECUTABLES/obbtool_intermediates/Main.o] Error 1
> 
> >>
> 
> >>
> 
> >>
> 
> >> Regards,
> 
> >>
> 
> >>
> 
> >>
> 
> >> Christopher De Cairos
> 
> >>
> 
> >> Integration Engineer - Webmaker
> 
> >>
> 
> >> Mozilla Foundation
> 
> >>
> 
> >>
> 
> >>
> 
> >> On 2013-10-16 12:33 PM, James Kruth wrote:
> 
> >>
> 
> >>> Chris,
> 
> >>> Let's see if it works for Christopher.  If it does, it would be 
> >>> incredibly helpful if you'd update MDN.  I'd eventually like to see this 
> >>> integrated into the Mac OS bootstrap script and the build file, but I'm 
> >>> not sure how long that may take.
> 
> >>> - James
> 
> >>> On Wednesday, October 16, 2013 8:36:45 AM UTC-4, Chris Mills wrote:
> 
> >>>> Ack, I guess this means
> 
> >>>> https://developer.mozilla.org/en-US/docs/Mozilla/Firefox_OS/Building#clang_errors_when_building_with_XCode_5_on_Mac
> 
> >>>> needs an update ;-)
> 
> >>>> Chris Mills
> 
> >>>>      Senior tech writer || Mozilla
> 
> >>>> developer.mozilla.org || MDN
> 
> >>>>      [email protected] || @chrisdavidmills
> 
> >>>> On 16 Oct 2013, at 07:19, James Kruth <[email protected]> wrote:
> 
> >>>>> What's the output of the following command?
> 
> >>>>> ls -l `which llvm-gcc`
> 
> >>>>> I've become convinced that depending on llvm-gcc is the wrong way to 
> >>>>> solve this problem, as it's only going to work on systems that once had 
> >>>>> Xcode 4 installed.  I have another solution that seems much more 
> >>>>> robust, and is working (inasmuch as the build has not failed).  Here 
> >>>>> are the steps:
> 
> >>>>> 1. Start with a clean tree configured for inari (I've been running a 
> >>>>> ./build.sh clean before each test).  Make sure you don't have a 
> >>>>> .userconfig.
> 
> >>>>> 2. Run the following commands:
> 
> >>>>>    a. brew uninstall gcc-4.6
> 
> >>>>>    b. brew install --enable-cxx 
> >>>>> https://gist.github.com/artlogic/6988658/raw/aeb9d1ea098274ad3f3fe2637b9df7f308a8a120/gcc-4.6.rb
> 
> >>>>> The above is a minor modification of the b2g gcc-4.6 formula to enable 
> >>>>> multilibs.
> 
> >>>>> 3. Change HOST_CXX in build/core/combo/HOST_darwin-x86.mk to |g++-4.6|. 
> >>>>>  I've actually modified mine a bit more so it works like the gcc-4.6 
> >>>>> assignment above.  Here's a diff:
> 
> >>>>> +HOST_CXX := g++-4.6
> 
> >>>>> +ifeq (,$(wildcard /usr/local/bin/g++-4.6))
> 
> >>>>> HOST_CXX := g++
> 
> >>>>> +endif
> 
> >>>>> After this, you should be able to build as normal.  This should work on 
> >>>>> ANY system, not just systems that have a stray llvm-gcc on them from 
> >>>>> Xcode 4.  You've already gone above and beyond in solving this problem. 
> >>>>>  If you have additional time, I'd appreciate your feedback on this 
> >>>>> technique.
> 
> >>>>> As an aside, I'm still having trouble flashing the images, but I 
> >>>>> believe that comes down to needing an updated boot.img, which is talked 
> >>>>> about here: 
> >>>>> http://sl.edujose.org/2013/10/adapted-boot-image-for-use-with-b2g.html
> 
> >>>>> - James
> 
> >>>>> On Tuesday, October 15, 2013 8:42:47 PM UTC-4, Christopher De Cairos 
> >>>>> wrote:
> 
> >>>>>> I've attempted rebuilding from scratch, changing only the HOST_CXX var
> 
> >>>>>> generated in build/core/combo/HOST_darwin-x86.mk to llvm-gcc
> 
> >>>>>> I get this build error:
> 
> >>>>>> configure: error: Your host toolchain does not support C++0x/C++11 
> >>>>>> mode properly. Please upgrade your toolchain
> 
> >>>>>> *** Fix above errors and then restart with               "make -f 
> >>>>>> client.mk build"
> 
> >>>>>> make[3]: *** [configure] Error 1
> 
> >>>>>> make[2]: *** [/Volumes/firefoxos/B2G/objdir-gecko/Makefile] Error 2
> 
> >>>>>> make[1]: *** [build] Error 2
> 
> >>>>>> make: *** 
> >>>>>> [out/target/product/inari/obj/DATA/gecko_intermediates/gecko] Error 2
> 
> >>>>>> which I haven't seen before. not sure what's causing it.
> 
> >>>>>> Regards,
> 
> >>>>>> Christopher De Cairos
> 
> >>>>>> Integration Engineer - Webmaker
> 
> >>>>>> Mozilla Foundation
> 
> >>>>>> On 2013-10-15 12:16 PM, James Kruth wrote:
> 
> >>>>>>> Christopher,
> 
> >>>>>>> I've been able to make (what I think) is a good build simply by 
> >>>>>>> changing HOST_CXX inside build/core/combo/HOST_darwin-x86.mk to 
> >>>>>>> |llvm-g++|.  None of the other changes appeared to be necessary for 
> >>>>>>> me.  I wonder if you might confirm this on your end?
> 
> >>>>>>> I haven't been able to flash this build, I think because no boot.img 
> >>>>>>> is being created, and updates need to be made.  Are you flashing on a 
> >>>>>>> production device?  If so, did you run into problems here?  If not, 
> >>>>>>> then it could be something is failing because I didn't attempt to 
> >>>>>>> make the other changes you mentioned.
> 
> >>>>>>> - James
> 
> >>>>>>> On Tuesday, October 15, 2013 9:58:28 AM UTC-4, Christopher De Cairos 
> >>>>>>> wrote:
> 
> >>>>>>>> I've made a few minor changes to the MDN article to fix some errors I
> 
> >>>>>>>> made in my original description. Thanks for documenting this!
> 
> >>>>>>>> Regards,
> 
> >>>>>>>> Christopher De Cairos
> 
> >>>>>>>> Integration Engineer - Webmaker
> 
> >>>>>>>> Mozilla Foundation
> 
> >>>>>>>> On 2013-10-15 3:52 AM, Chris Mills wrote:
> 
> >>>>>>>>> Hi is great Christopher - thanks! I've added the following entry to 
> >>>>>>>>> the documentation:
> 
> >>>>>>>>> https://developer.mozilla.org/en-US/docs/Mozilla/Firefox_OS/Building#clang_errors_when_building_with_XCode_5_on_Mac
> 
> >>>>>>>>> Let me know if you think this reads ok.
> 
> >>>>>>>>> Chris Mills
> 
> >>>>>>>>>       Senior tech writer || Mozilla
> 
> >>>>>>>>> developer.mozilla.org || MDN
> 
> >>>>>>>>>       [email protected] || @chrisdavidmills
> 
> >>>>>>>>> On 14 Oct 2013, at 17:53, Christopher De Cairos 
> >>>>>>>>> <[email protected]> wrote:
> 
> >>>>>>>>>> I'll do my best to describe the problem below:
> 
> >>>>>>>>>> While building B2G with it configured for Inari, I ran into 
> >>>>>>>>>> trouble on OSX. Xcode 5 changes the gcc and g++ compilers in 
> >>>>>>>>>> /usr/bin, which breaks the build process if you try to use them to 
> >>>>>>>>>> compile.
> 
> >>>>>>>>>> In order to work around the problem, I decided to manually change 
> >>>>>>>>>> instances of `g++` and `gcc`  to `llvm-g++-4.2` and `llvm-ggc-4.2` 
> >>>>>>>>>>  in the following folders:
> 
> >>>>>>>>>> objdir-gecko/_virtualenv/lib/python2.7/config/Makefile
> 
> >>>>>>>>>>    � CC
> 
> >>>>>>>>>>    � CXX
> 
> >>>>>>>>>> build/core/combo/HOST_darwin-x86.mk
> 
> >>>>>>>>>>    � HOST_OBJCC
> 
> >>>>>>>>>>    � HOST_CC
> 
> >>>>>>>>>>    � HOST_CXX
> 
> >>>>>>>>>> I also added this to my .userconfig file, for good measure:
> 
> >>>>>>>>>> export CXX=llvm-g++-4.6
> 
> >>>>>>>>>> export CC=llvm-gcc-4.6
> 
> >>>>>>>>>> export CPP=llvm-gcc-4.6
> 
> >>>>>>>>>> Let me know if I can clarify any of the above a bit further.
> 
> >>>>>>>>>> Regards,
> 
> >>>>>>>>>> Christopher De Cairos
> 
> >>>>>>>>>> Integration Engineer - Webmaker
> 
> >>>>>>>>>> Mozilla Foundation
> 
> >>>>>>>>>> On 2013-10-14 1:25 AM, Chris Mills wrote:
> 
> >>>>>>>>>>> And a clear description of the problem would be good too. I'll 
> >>>>>>>>>>> then get it added to MDN in an appropriate place.
> 
> >>>>>>>>>>> cheers!
> 
> >>>>>>>>>>> Chris Mills
> 
> >>>>>>>>>>>       Senior tech writer || Mozilla
> 
> >>>>>>>>>>> developer.mozilla.org || MDN
> 
> >>>>>>>>>>> [email protected]
> 
> >>>>>>>>>>>     || @chrisdavidmills
> 
> >>>>>>>>>>> On 14 Oct 2013, at 01:02, James Kruth
> 
> >>>>>>>>>>> <[email protected]>
> 
> >>>>>>>>>>>     wrote:
> 
> >>>>>>>>>>>> Christopher,
> 
> >>>>>>>>>>>> That's great news!  I'm wondering if you might be so kind as to 
> >>>>>>>>>>>> outline the files you needed to change?
> 
> >>>>>>>>>>>> - James
> 
> >>>>>>>>>>>> On Sunday, October 13, 2013 6:40:09 PM UTC-4, Christopher De 
> >>>>>>>>>>>> Cairos wrote:
> 
> >>>>>>>>>>>>> James,
> 
> >>>>>>>>>>>>> Yes, the changes I made allowed me to successfully build B2G 
> >>>>>>>>>>>>> v1.3. There
> 
> >>>>>>>>>>>>> were a few other bumps down the road - some other spots needed 
> >>>>>>>>>>>>> updates
> 
> >>>>>>>>>>>>> for CC and CXX vars
> 
> >>>>>>>>>>>>> In the end, I was able to successfully flash it onto my ZTE 
> >>>>>>>>>>>>> Open, and so
> 
> >>>>>>>>>>>>> far it's working great.
> 
> >>>>>>>>>>>>> Your assessment of the Xcode 5 update sounds about right to me, 
> >>>>>>>>>>>>> gcc and
> 
> >>>>>>>>>>>>> g++ on my machine were no longer symlinked to my llvm-* 
> >>>>>>>>>>>>> binaries as they
> 
> >>>>>>>>>>>>> may have been in Xcode 4.x
> 
> >>>>>>>>>>>>> If the output from the build process was correct, then I won't 
> >>>>>>>>>>>>> be able
> 
> >>>>>>>>>>>>> to build for a simulator, but I haven't tried that.
> 
> >>>>>>>>>>>>> Regards,
> 
> >>>>>>>>>>>>> Christopher De Cairos
> 
> >>>>>>>>>>>>> Integration Engineer - Webmaker
> 
> >>>>>>>>>>>>> Mozilla Foundation
> 
> >>>>>>>>>>>>> On 2013-10-12 4:12 PM, James Kruth wrote:
> 
> >>>>>>>>>>>>>> Some updates on what appears to be going on here:
> 
> >>>>>>>>>>>>>> * On Xcode 4.x, /usr/bin/gcc was a symlink, likely to 
> >>>>>>>>>>>>>> llvm-gcc-4.2 (which apparently wouldn't produce a usable 
> >>>>>>>>>>>>>> emulator build - I'm not sure about inari)
> 
> >>>>>>>>>>>>>> * With the Xcode 5 update, the CLI tools have replaced that 
> >>>>>>>>>>>>>> symlink with an actual binary, which is clang, as opposed to 
> >>>>>>>>>>>>>> gcc.
> 
> >>>>>>>>>>>>>> The build obviously doesn't compile with clang, and so things 
> >>>>>>>>>>>>>> are broken in their current state.  I'd be curious to know if 
> >>>>>>>>>>>>>> Christopher's modifications gave him a working inari build.
> 
> >>>>>>>>>>>>>> Where is the appropriate place to report a bug like this?
> 
> >>>>>>>>>>>>>> - James
> 
> >>>>>>>>>>>>>> On Saturday, October 12, 2013 12:59:23 AM UTC-4, James Kruth 
> >>>>>>>>>>>>>> wrote:
> 
> >>>>>>>>>>>>>>> Actually, Keon is called out specifically:
> 
> >>>>>>>>>>>>>>> https://developer.mozilla.org/en-US/docs/Mozilla/Firefox_OS/Firefox_OS_build_prerequisites#Requirements_for_Mac_OS_X
> 
> >>>>>>>>>>>>>>> At this point, I'm willing to say it's not a problem with 
> >>>>>>>>>>>>>>> inari, as I just attempted to build for emulator and received 
> >>>>>>>>>>>>>>> the same clang errors.  There's definitely a problem with the 
> >>>>>>>>>>>>>>> build process.
> 
> >>>>>>>>>>>>>>> - James
> 
> >>>>>>>>>>>>>>> On Saturday, October 12, 2013 12:55:27 AM UTC-4, Caio Lima 
> >>>>>>>>>>>>>>> wrote:
> 
> >>>>>>>>>>>>>>>> Guys, I don't remember where, but I have seen in 
> >>>>>>>>>>>>>>>> documentation that the
> 
> >>>>>>>>>>>>>>>> inari build is not working on Mac OS X. I guess it was in 
> >>>>>>>>>>>>>>>> "Build the
> 
> >>>>>>>>>>>>>>>> Firefox OS" on MDN
> 
> >>>>>>>>>>>>>>>> Give a look on this tutorial.
> 
> >>>>>>>>>>>>>>>> On Saturday, October 12, 2013, Reuben Morais wrote:
> 
> >>>>>>>>>>>>>>>>> And yes, I'm pretty sure the bootstrap script should be 
> >>>>>>>>>>>>>>>>> passing
> 
> >>>>>>>>>>>>>>>>> --enable-cxx to GCC's configure script. Not entirely sure 
> >>>>>>>>>>>>>>>>> because it looks
> 
> >>>>>>>>>>>>>>>>> like Clang is being picked up (and causing the errors in 
> >>>>>>>>>>>>>>>>> the OP). You can
> 
> >>>>>>>>>>>>>>>>> manually fix it by doing:
> 
> >>>>>>>>>>>>>>>>>      brew install
> 
> >>>>>>>>>>>>>>>>> https://raw.github.com/mozilla-b2g/B2G/master/scripts/homebrew/gcc-4.6.rb--enable-cxx
> 
> >>>>>>>>>>>>>>>>> -- reuben
> 
> >>>>>>>>>>>>>>>> -- 
> 
> >>>>>>>>>>>>>>>> Caio Lima
> 
> >>>>>>>>>>>>>> _______________________________________________
> 
> >>>>>>>>>>>>>> dev-b2g mailing list
> 
> >>>>>>>>>>>>>> [email protected]
> 
> >>>>>>>>>>>>>> https://lists.mozilla.org/listinfo/dev-b2g
> 
> >>>>>>>>>>>> _______________________________________________
> 
> >>>>>>>>>>>> dev-b2g mailing list
> 
> >>>>>>>>>>>> [email protected]
> 
> >>>>>>>>>>>> https://lists.mozilla.org/listinfo/dev-b2g
> 
> >>>>>>>>>>> _______________________________________________
> 
> >>>>>>>>>>> dev-b2g mailing list
> 
> >>>>>>>>>>> [email protected]
> 
> >>>>>>>>>>> https://lists.mozilla.org/listinfo/dev-b2g
> 
> >>>>>>> _______________________________________________
> 
> >>>>>>> dev-b2g mailing list
> 
> >>>>>>> [email protected]
> 
> >>>>>>> https://lists.mozilla.org/listinfo/dev-b2g
> 
> >>>>> _______________________________________________
> 
> >>>>> dev-b2g mailing list
> 
> >>>>> [email protected]
> 
> >>>>> https://lists.mozilla.org/listinfo/dev-b2g
> 
> >>> _______________________________________________
> 
> >>> dev-b2g mailing list
> 
> >>> [email protected]
> 
> >>> https://lists.mozilla.org/listinfo/dev-b2g
> 
> > _______________________________________________
> 
> > dev-b2g mailing list
> 
> > [email protected]
> 
> > https://lists.mozilla.org/listinfo/dev-b2g
_______________________________________________
dev-b2g mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-b2g

Reply via email to