Re: Safe-5.3 failed to be built on Mountain Lion (OS X 10.8)

[John H Palmieri]

On Tuesday, September 25, 2012 1:57:42 PM UTC-7, JBT wrote:

Hi,

I tried to build sage from source code on my mac running Mountain lion (10.8.2), but didn't succeed.

Sage won't build, as is, on Mountain Lion. Several of us have had good luck with the following procedure:

- make sure you've updated to 10.8.2 (which you have).
- make sure you've updated to the latest version of Xcode (4.5), and install the command line tools.
- get the scipy spkg from the link given at http://trac.sagemath.org/sage_trac/ticket/13309, and put it in SAGE_ROOT/spkg/standard.
- run 'make'.

Good luck!

--
John

[JBT]

Hi, John,

Thanks for the tip. I tried it, but no luck :-( The 'make' got stuck pretty much at the same place.

JBT

[Michael Welsh]

On 26/09/2012, at 9:45 AM, JBT <jianb...@gmail.com> wrote:
> Hi, John,
>
> Thanks for the tip. I tried it, but no luck :-( The 'make' got stuck pretty
> much at the same place.

1. Did you try again from scratch? Either by running 'make clean' or unpacking the source again.
2. Did you delete the older scipy spkg?

Michael

[JBT]

Michael,

OK, I didn't try either of the two things last time. But I didn't get lucky either after 'make clean' and deleting old scipy package. It failed at the same spot.

JBT

[John H Palmieri]

On Tuesday, September 25, 2012 3:11:52 PM UTC-7, JBT wrote:

Michael,

OK, I didn't try either of the two things last time. But I didn't get lucky either after 'make clean' and deleting old scipy package. It failed at the same spot.

JBT

On Tuesday, September 25, 2012 3:02:30 PM UTC-7, yomcat wrote:

On 26/09/2012, at 9:45 AM, JBT <jianb...@gmail.com> wrote:
> Hi, John,
>
> Thanks for the tip. I tried it, but no luck :-( The 'make' got stuck pretty
> much at the same place.

1. Did you try again from scratch? Either by running 'make clean' or unpacking the source again.
2. Did you delete the older scipy spkg?

'make clean' doesn't do very much. 'make distclean' is more effective. I don't know if it will help, though.

What does 'uname -a' return? What sort of machine is this?

--
John
 

[JBT]

John,

FYI, here is the uname info:

>$ uname -a

Darwin jbtlap.ssl.berkeley.edu 12.2.0 Darwin Kernel Version 12.2.0: Sat Aug 25 00:48:52 PDT 2012; root:xnu-2050.18.24~1/RELEASE_X86_64 x86_64

[John H Palmieri]

On Tuesday, September 25, 2012 3:24:08 PM UTC-7, JBT wrote:

John,

FYI, here is the uname info:

>$ uname -a

Darwin jbtlap.ssl.berkeley.edu 12.2.0 Darwin Kernel Version 12.2.0: Sat Aug 25 00:48:52 PDT 2012; root:xnu-2050.18.24~1/RELEASE_X86_64 x86_64

From your log:

   gcc ... -arch i386 -arch x86_64 ...
   gcc: error: i386: No such file or directory
   gcc: error: x86_64: No such file or directory

This looks suspicious. Did Sage's gcc spkg build? Or have you installed gcc yourself? If you've installed it yourself, what does "gcc --version" say? Also if you've installed it yourself, start with a clean Sage tarball, get the new scipy spkg (although this won't be relevant if the python package fails to build) and do this:

   $ export SAGE_INSTALL_GCC=yes
   $ make

Does that help?

--
John

[JBT]

Hi John,

Thanks for following up on this. So, I tried your suggestion. Here are a few shell lines:

>$ make

cd spkg && \

"../spkg/pipestatus" \

"env SAGE_PARALLEL_SPKG_BUILD='' ./install all 2>&1" \

"tee -a ../install.log"

Installing GCC because SAGE_INSTALL_GCC is set to 'yes'.

Unfortunately, still no luck. The building process stopped at the same place. Bummer :-(

Jianbao

[John H Palmieri]

On Thursday, September 27, 2012 3:16:07 PM UTC-7, JBT wrote:

Hi John,

Thanks for following up on this. So, I tried your suggestion. Here are a few shell lines:

>$ make

cd spkg && \

"../spkg/pipestatus" \

"env SAGE_PARALLEL_SPKG_BUILD='' ./install all 2>&1" \

"tee -a ../install.log"

Installing GCC because SAGE_INSTALL_GCC is set to 'yes'.

Unfortunately, still no luck. The building process stopped at the same place. Bummer :-(

You didn't answer the other questions: have you installed another version of gcc? What does "gcc --version" say? Oh, and also, what does "command -v gcc" say?

--
John