Re: [sage-devel] ccache (GAP purposely unsets CC which screws up Sun Studio build.)

[William Stein]

On Sun, Sep 27, 2009 at 9:31 AM, Dr. David Kirkby
<david....@onetel.net> wrote:
>
>
>
> The spkg-install for gap-4.4.10.p12 says
>
> echo "*WARNING*: Unsetting CC since that tends to break GAP building"
> unset CC echo "*WARNING*: Unsetting CXX since that tends to break GAP
> building" unset CXX
>
> This appears to be done by Michael Abshoff as SPKG.txt shows:
> gap-4.4.10.p8 (Michael Abshoff, June 16th, 2008) ¶
>
>     * unset CC in spkg-install (work around for #2575)
>
> Looking at http://sagetrac.org/sage_trac/ticket/2575 I can't help feel
> there must be a better solution.

Find it! :-)

But seriously, this ccache looks really useful for the Cygwin port,
where building using GCC is painfully slow (we're talking 1-2 days to
build sage from source on a 2.6Ghz box).

The ccache page says "ccache is a compiler cache. It acts as a caching
pre-processor to C/C++ compilers, using the -E compiler switch and a
hash to detect when a compilation can be satisfied from cache. This
often results in a 5 to 10 times speedup in common compilations."

William

[William Stein]

On Sun, Sep 27, 2009 at 12:48 PM, Dr. David Kirkby
<david....@onetel.net> wrote:

>
> William Stein wrote:
>> On Sun, Sep 27, 2009 at 9:31 AM, Dr. David Kirkby
>> <david....@onetel.net> wrote:
>>>
>>>
>>> The spkg-install for gap-4.4.10.p12 says
>>>
>>> echo "*WARNING*: Unsetting CC since that tends to break GAP building"
>>> unset CC echo "*WARNING*: Unsetting CXX since that tends to break GAP
>>> building" unset CXX
>>>
>>> This appears to be done by Michael Abshoff as SPKG.txt shows:
>>> gap-4.4.10.p8 (Michael Abshoff, June 16th, 2008) ¶
>>>
>>>     * unset CC in spkg-install (work around for #2575)
>>>
>>> Looking at http://sagetrac.org/sage_trac/ticket/2575 I can't help feel
>>> there must be a better solution.
>>
>> Find it!  :-)
>>
>> But seriously, this ccache looks really useful for the Cygwin port,
>> where building using GCC is painfully slow (we're talking 1-2 days to
>> build sage from source on a 2.6Ghz box).
>
>

> I thought the cygwin port was basically dead, and one was supposed to
> use Virtualbox or similar.
>
> Is that not correct?

That couldn't be more wrong. In fact, I will bet you Sage will fully
support Cygwin before Solaris is supported (meant in the spirit of
friendly competition).

You might want to subscribe to

http://groups.google.com/group/sage-windows

and see

http://trac.sagemath.org/sage_trac/ticket/6743

I have a VirtualBox install of Windows XP and cygwin on
bsd.math.washington.edu, which I can ssh into, and several sage
developers have accounts on it.

In fact every component of Sage builds (with some work) under Cygwin
now except R, but the Sage library doesn't import yet. The current
thing that has me stumped is this:

wstein@winxp ~/sage-4.1.2.alpha1/devel/x$ ../../sage -ipython
Python 2.6.2 (r262:71600, Sep 20 2009, 20:36:28)
Type "copyright", "credits" or "license" for more information.
In [1]: import sage.all
---------------------------------------------------------------------------
ImportError Traceback (most recent call last)
/home/wstein/sage-4.1.2.alpha1/devel/x/linbox.pxd in
sage.matrix.matrix_integer_dense
(sage/matrix/matrix_integer_dense.c:39095)()
ImportError: Bad address
-------------------------------------------------------

I think this is an issue with how the Linbox wrapper was linked (using
ld instead of gcc), but I haven't got around to fixing it yet. It's
the current problem though.

-- William