[gentoo-science] [sage-on-gentoo] cannot build Sage because of Sympy-0.7.1

[gentoo-science] [sage-on-gentoo] cannot build Sage because of Sympy-0.7.1

[v_2e]

  Hello!
  I've noticed that 'lcalc' is available again and tried to emerge
Sage. This time it failed because of 'dev-python/sympy-0.7.1'. It
refuses to be built and gives the following error:


>>> Compiling source
>>> in /var/tmp/portage/dev-python/sympy-0.7.1/work/sympy-0.7.1 ...
 * Building of dev-python/sympy-0.7.1 with CPython 2.7...
python2.7 setup.py build -b build-2.7
Setting permissions of DOT_SAGE directory so only you can read and
write it. 
Traceback (most recent call last):
  File "setup.py", line 34, in <module>
    import sympy
  File
"/var/tmp/portage/dev-python/sympy-0.7.1/work/sympy-0.7.1/sympy/__init__.py",
line 20, in <module> from sympy.core import * 
File
"/var/tmp/portage/dev-python/sympy-0.7.1/work/sympy-0.7.1/sympy/core/__init__.py",
line 8, in <module> from expr import Expr, AtomicExpr 
File
"/var/tmp/portage/dev-python/sympy-0.7.1/work/sympy-0.7.1/sympy/core/expr.py",
line 4, in <module> from evalf import EvalfMixin 
File
"/var/tmp/portage/dev-python/sympy-0.7.1/work/sympy-0.7.1/sympy/core/evalf.py",
line 6, in <module> import mpmath.libmp as libmp 
File
"/usr/lib64/python2.7/site-packages/mpmath/__init__.py", line 10, in
<module> mp = MPContext() 
File
"/usr/lib64/python2.7/site-packages/mpmath/ctx_mp.py", line 72, in
__init__ ctx.init_builtins() 
File
"/usr/lib64/python2.7/site-packages/mpmath/ctx_mp.py", line 101, in
init_builtins ctx.one = ctx.make_mpf(fone) 
File "ext_main.pyx", line 493, in
sage.libs.mpmath.ext_main.Context.make_mpf
(sage/libs/mpmath/ext_main.c:6425) 
File "ext_main.pyx", line 503, in
sage.libs.mpmath.ext_main.Context.make_mpf
(sage/libs/mpmath/ext_main.c:6360) 
File "ext_impl.pyx", line 268, in
sage.libs.mpmath.ext_impl.MPF_set_tuple
(sage/libs/mpmath/ext_impl.c:3845) 
File "integer.pyx", line 752, in
sage.rings.integer.Integer.__init__ (sage/rings/integer.c:7812)
TypeError: unable to coerce <type 'mpz'> to an integer



  Regards,
    Vladimir

----- 
 <v_2e@ukr.net>

Re: [gentoo-science] [sage-on-gentoo] cannot build Sage because of Sympy-0.7.1

[fbissey]

Quoting v_2e@ukr.net:

>   Hello!
>   I've noticed that 'lcalc' is available again and tried to emerge
> Sage. This time it failed because of 'dev-python/sympy-0.7.1'. It
> refuses to be built and gives the following error:
>
>
>>>> Compiling source
>>>> in /var/tmp/portage/dev-python/sympy-0.7.1/work/sympy-0.7.1 ...
>  * Building of dev-python/sympy-0.7.1 with CPython 2.7...
> python2.7 setup.py build -b build-2.7
> Setting permissions of DOT_SAGE directory so only you can read and
> write it.
> Traceback (most recent call last):
>   File "setup.py", line 34, in <module>
>     import sympy
>   File
> "/var/tmp/portage/dev-python/sympy-0.7.1/work/sympy-0.7.1/sympy/__init__.py",
> line 20, in <module> from sympy.core import *
> File
> "/var/tmp/portage/dev-python/sympy-0.7.1/work/sympy-0.7.1/sympy/core/__init__.py",
> line 8, in <module> from expr import Expr, AtomicExpr
> File
> "/var/tmp/portage/dev-python/sympy-0.7.1/work/sympy-0.7.1/sympy/core/expr.py",
> line 4, in <module> from evalf import EvalfMixin
> File
> "/var/tmp/portage/dev-python/sympy-0.7.1/work/sympy-0.7.1/sympy/core/evalf.py",
> line 6, in <module> import mpmath.libmp as libmp
> File
> "/usr/lib64/python2.7/site-packages/mpmath/__init__.py", line 10, in
> <module> mp = MPContext()
> File
> "/usr/lib64/python2.7/site-packages/mpmath/ctx_mp.py", line 72, in
> __init__ ctx.init_builtins()
> File
> "/usr/lib64/python2.7/site-packages/mpmath/ctx_mp.py", line 101, in
> init_builtins ctx.one = ctx.make_mpf(fone)
> File "ext_main.pyx", line 493, in
> sage.libs.mpmath.ext_main.Context.make_mpf
> (sage/libs/mpmath/ext_main.c:6425)
> File "ext_main.pyx", line 503, in
> sage.libs.mpmath.ext_main.Context.make_mpf
> (sage/libs/mpmath/ext_main.c:6360)
> File "ext_impl.pyx", line 268, in
> sage.libs.mpmath.ext_impl.MPF_set_tuple
> (sage/libs/mpmath/ext_impl.c:3845)
> File "integer.pyx", line 752, in
> sage.rings.integer.Integer.__init__ (sage/rings/integer.c:7812)
> TypeError: unable to coerce <type 'mpz'> to an integer
>
>
Hi Vladimir,

I cannot reproduce that. Which version of sage do you have currently installed
and was any component updated recently (mpmath,mpfr,gmp?).

Francois

Re: [gentoo-science] [sage-on-gentoo] cannot build Sage because of Sympy-0.7.1

[v_2e]

  Hello!

On Wed, 15 Aug 2012 09:31:12 +1200
fbissey@slingshot.co.nz wrote:

> Quoting v_2e@ukr.net:
> 
> >   Hello!
> >   I've noticed that 'lcalc' is available again and tried to emerge
> > Sage. This time it failed because of 'dev-python/sympy-0.7.1'. It
> > refuses to be built and gives the following error:
> >
> > . . .
> >
> Hi Vladimir,
> 
> I cannot reproduce that. Which version of sage do you have currently
> installed and was any component updated recently (mpmath,mpfr,gmp?).
> 
  Well, actually I lost my "/var" partition several days ago, so now I'm
re-emerging everything although I have already got this "everything" on
the "/usr" partition.
  So when I run 'emerge sage', it pulls 'dev-python/sympy-0.7.1' as a
dependency. That is why I'm not sure I can understand why the SymPy
build log mentions Sage at all. How can it know about Sage if Sage is
intended to be merged *after* sympy?
  I'm trying to emerge Sage 5.2-r2, and I have Sage-5.2 (I don't
remember exactly whether it was '-r2' or '-r1' though) installed on my
system.

  Thanks!
    Vladimir


----- 
 <v_2e@ukr.net>

Re: [gentoo-science] [sage-on-gentoo] cannot build Sage because of Sympy-0.7.1

[fbissey]

Quoting v_2e@ukr.net:

>   Hello!
>
> On Wed, 15 Aug 2012 09:31:12 +1200
> fbissey@slingshot.co.nz wrote:
>
>> Quoting v_2e@ukr.net:
>>
>> >   Hello!
>> >   I've noticed that 'lcalc' is available again and tried to emerge
>> > Sage. This time it failed because of 'dev-python/sympy-0.7.1'. It
>> > refuses to be built and gives the following error:
>> >
>> > . . .
>> >
>> Hi Vladimir,
>>
>> I cannot reproduce that. Which version of sage do you have currently
>> installed and was any component updated recently (mpmath,mpfr,gmp?).
>>
>   Well, actually I lost my "/var" partition several days ago, so now I'm
> re-emerging everything although I have already got this "everything" on
> the "/usr" partition.
>   So when I run 'emerge sage', it pulls 'dev-python/sympy-0.7.1' as a
> dependency. That is why I'm not sure I can understand why the SymPy
> build log mentions Sage at all. How can it know about Sage if Sage is
> intended to be merged *after* sympy?
>   I'm trying to emerge Sage 5.2-r2, and I have Sage-5.2 (I don't
> remember exactly whether it was '-r2' or '-r1' though) installed on my
> system.
>
Actually mpmath and sympy autodetect sage at run-time. If it is present they
will use it. There is a variable to set off that behavior in mpmath but it the
mechanism is currently broken. I talked to both upstream about it and 
hopefully
something will be done about it eventually.

Technically sympy is a run time dependency of sage so you can emerge it
afterwards. Two solutions: skip it and come back to it later when sage 
has been
merged again and hopefully whatever is currently broken (and it is 
broken) will
be fixed. The other solution is to unmerge sage and emerge sympy then 
start the
merge process of sage again.

Francois

Re: [gentoo-science] [sage-on-gentoo] cannot build Sage because of Sympy-0.7.1

[v_2e]

  Hello!

On Wed, 15 Aug 2012 13:32:46 +1200
fbissey@slingshot.co.nz wrote:
>  ...
> Actually mpmath and sympy autodetect sage at run-time. If it is
> present they will use it. There is a variable to set off that
> behavior in mpmath but it the mechanism is currently broken. I talked
> to both upstream about it and hopefully
> something will be done about it eventually.
> 
> Technically sympy is a run time dependency of sage so you can emerge
> it afterwards. Two solutions: skip it and come back to it later when
> sage has been
> merged again and hopefully whatever is currently broken (and it is 
> broken) will
> be fixed. The other solution is to unmerge sage and emerge sympy then 
> start the
> merge process of sage again.
> 
  I tried different things including:

  emerge -va --nodep sage               # success
  emerge -vCa sage               # success
  emerge -1va sympy               # success
  emerge -va sage               # success
  emerge -1va sympy               # fails with the same error as before


    Regards,
        Vladimir

----- 
 <v_2e@ukr.net>

Re: [gentoo-science] [sage-on-gentoo] cannot build Sage because of Sympy-0.7.1

[Francois Bissey]

On Wed, 15 Aug 2012 13:32:51 v_2e@ukr.net wrote:
>   Hello!
> 
> On Wed, 15 Aug 2012 13:32:46 +1200
> 
> fbissey@slingshot.co.nz wrote:
> >  ...
> > 
> > Actually mpmath and sympy autodetect sage at run-time. If it is
> > present they will use it. There is a variable to set off that
> > behavior in mpmath but it the mechanism is currently broken. I talked
> > to both upstream about it and hopefully
> > something will be done about it eventually.
> > 
> > Technically sympy is a run time dependency of sage so you can emerge
> > it afterwards. Two solutions: skip it and come back to it later when
> > sage has been
> > merged again and hopefully whatever is currently broken (and it is
> > broken) will
> > be fixed. The other solution is to unmerge sage and emerge sympy then
> > start the
> > merge process of sage again.
> 
>   I tried different things including:
> 
>   emerge -va --nodep sage               # success
>   emerge -vCa sage               # success
>   emerge -1va sympy               # success
>   emerge -va sage               # success
>   emerge -1va sympy               # fails with the same error as before
> 
> 
hum...... what does 
printenv | grep SAGE
return?

Francois

Re: [gentoo-science] [sage-on-gentoo] cannot build Sage because of Sympy-0.7.1

[v_2e]

On Wed, 15 Aug 2012 22:36:35 +1200
Francois Bissey <fbissey@slingshot.co.nz> wrote:

> hum...... what does 
> printenv | grep SAGE
> return?
  Nothing. I checked both in my user's terminal and in the root's one.


    Regards,
        Vladimir

----- 
 <v_2e@ukr.net>

Re: [gentoo-science] [sage-on-gentoo] cannot build Sage because of Sympy-0.7.1

[fbissey]

Quoting v_2e@ukr.net:

> On Wed, 15 Aug 2012 22:36:35 +1200
> Francois Bissey <fbissey@slingshot.co.nz> wrote:
>
>> hum...... what does printenv | grep SAGE
>> return?
>   Nothing. I checked both in my user's terminal and in the root's one.
>
I was thinking out loud. Yes you wouldn't have anything if you are at sage-5.2
and that's not a recent upgrade. From a python (2.7) shell, what 
happens if you
do
import mpmath

import sympy

?
Also do you you have any problem emerging mpmath.

Francois

Re: [gentoo-science] [sage-on-gentoo] cannot build Sage because of Sympy-0.7.1

[v_2e]

  Hello!

On Thu, 16 Aug 2012 10:03:16 +1200
fbissey@slingshot.co.nz wrote:

> From a python (2.7) shell, what happens if you do
> 
> import mpmath
> 

>>> import mpmath
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib64/python2.7/site-packages/mpmath/__init__.py", line
10, in <module> mp = MPContext()
  File "/usr/lib64/python2.7/site-packages/mpmath/ctx_mp.py", line 72,
in __init__ ctx.init_builtins()
  File "/usr/lib64/python2.7/site-packages/mpmath/ctx_mp.py", line 101,
in init_builtins ctx.one = ctx.make_mpf(fone)
  File "ext_main.pyx", line 493, in
sage.libs.mpmath.ext_main.Context.make_mpf
(sage/libs/mpmath/ext_main.c:6425) 
  File "ext_main.pyx", line 503, in
sage.libs.mpmath.ext_main.Context.make_mpf
(sage/libs/mpmath/ext_main.c:6360) 
  File "ext_impl.pyx", line 268, in
sage.libs.mpmath.ext_impl.MPF_set_tuple
(sage/libs/mpmath/ext_impl.c:3845) 
  File "integer.pyx", line 752, in
sage.rings.integer.Integer.__init__ (sage/rings/integer.c:7812)
TypeError: unable to coerce <type 'mpz'> to an integer
>>> 



> import sympy
> 
>>> import sympy
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib64/python2.7/site-packages/sympy/__init__.py", line 20,
in <module> from sympy.core import *
  File "/usr/lib64/python2.7/site-packages/sympy/core/__init__.py",
line 8, in <module> from expr import Expr, AtomicExpr
  File "/usr/lib64/python2.7/site-packages/sympy/core/expr.py", line 4,
in <module> from evalf import EvalfMixin
  File "/usr/lib64/python2.7/site-packages/sympy/core/evalf.py", line
6, in <module> import mpmath.libmp as libmp
  File "/usr/lib64/python2.7/site-packages/mpmath/__init__.py", line
10, in <module> mp = MPContext()
  File "/usr/lib64/python2.7/site-packages/mpmath/ctx_mp.py", line 72,
in __init__ ctx.init_builtins()
  File "/usr/lib64/python2.7/site-packages/mpmath/ctx_mp.py", line 101,
in init_builtins ctx.one = ctx.make_mpf(fone)
  File "ext_main.pyx", line 493, in
sage.libs.mpmath.ext_main.Context.make_mpf
(sage/libs/mpmath/ext_main.c:6425) 
  File "ext_main.pyx", line 503, in
sage.libs.mpmath.ext_main.Context.make_mpf
(sage/libs/mpmath/ext_main.c:6360) 
  File "ext_impl.pyx", line 268, in
sage.libs.mpmath.ext_impl.MPF_set_tuple
(sage/libs/mpmath/ext_impl.c:3845) 
  File "integer.pyx", line 752, in
sage.rings.integer.Integer.__init__ (sage/rings/integer.c:7812)
TypeError: unable to coerce <type 'mpz'> to an integer



> ?
> Also do you you have any problem emerging mpmath.
  No, I do not have any problems emerging mpmath. At least emerge does
not report of any problems during installation.

    Regards,
        Vladimir


----- 
 <v_2e@ukr.net>

Re: [gentoo-science] [sage-on-gentoo] cannot build Sage because of Sympy-0.7.1

[Francois Bissey]

On Thu, 16 Aug 2012 11:05:36 v_2e@ukr.net wrote:
>   Hello!
> 
> On Thu, 16 Aug 2012 10:03:16 +1200
> 
> fbissey@slingshot.co.nz wrote:
> > From a python (2.7) shell, what happens if you do
> > 
> > import mpmath
> > 
> >>> import mpmath
> 
> Traceback (most recent call last):
>   File "<stdin>", line 1, in <module>
>   File "/usr/lib64/python2.7/site-packages/mpmath/__init__.py", line
> 10, in <module> mp = MPContext()
>   File "/usr/lib64/python2.7/site-packages/mpmath/ctx_mp.py", line 72,
> in __init__ ctx.init_builtins()
>   File "/usr/lib64/python2.7/site-packages/mpmath/ctx_mp.py", line 101,
> in init_builtins ctx.one = ctx.make_mpf(fone)
>   File "ext_main.pyx", line 493, in
> sage.libs.mpmath.ext_main.Context.make_mpf
> (sage/libs/mpmath/ext_main.c:6425)
>   File "ext_main.pyx", line 503, in
> sage.libs.mpmath.ext_main.Context.make_mpf
> (sage/libs/mpmath/ext_main.c:6360)
>   File "ext_impl.pyx", line 268, in
> sage.libs.mpmath.ext_impl.MPF_set_tuple
> (sage/libs/mpmath/ext_impl.c:3845)
>   File "integer.pyx", line 752, in
> sage.rings.integer.Integer.__init__ (sage/rings/integer.c:7812)
> TypeError: unable to coerce <type 'mpz'> to an integer
> 
> > import sympy
> > 
> >>> import sympy
> 
> Traceback (most recent call last):
>   File "<stdin>", line 1, in <module>
>   File "/usr/lib64/python2.7/site-packages/sympy/__init__.py", line 20,
> in <module> from sympy.core import *
>   File "/usr/lib64/python2.7/site-packages/sympy/core/__init__.py",
> line 8, in <module> from expr import Expr, AtomicExpr
>   File "/usr/lib64/python2.7/site-packages/sympy/core/expr.py", line 4,
> in <module> from evalf import EvalfMixin
>   File "/usr/lib64/python2.7/site-packages/sympy/core/evalf.py", line
> 6, in <module> import mpmath.libmp as libmp
>   File "/usr/lib64/python2.7/site-packages/mpmath/__init__.py", line
> 10, in <module> mp = MPContext()
>   File "/usr/lib64/python2.7/site-packages/mpmath/ctx_mp.py", line 72,
> in __init__ ctx.init_builtins()
>   File "/usr/lib64/python2.7/site-packages/mpmath/ctx_mp.py", line 101,
> in init_builtins ctx.one = ctx.make_mpf(fone)
>   File "ext_main.pyx", line 493, in
> sage.libs.mpmath.ext_main.Context.make_mpf
> (sage/libs/mpmath/ext_main.c:6425)
>   File "ext_main.pyx", line 503, in
> sage.libs.mpmath.ext_main.Context.make_mpf
> (sage/libs/mpmath/ext_main.c:6360)
>   File "ext_impl.pyx", line 268, in
> sage.libs.mpmath.ext_impl.MPF_set_tuple
> (sage/libs/mpmath/ext_impl.c:3845)
>   File "integer.pyx", line 752, in
> sage.rings.integer.Integer.__init__ (sage/rings/integer.c:7812)
> TypeError: unable to coerce <type 'mpz'> to an integer
> 
> > ?
> > Also do you you have any problem emerging mpmath.
> 
>   No, I do not have any problems emerging mpmath. At least emerge does
> not report of any problems during installation.
> 
you have numpy-1.6.xx installed don't you? I am adding that to the list of 
pains related to the fact we are stuck with numpy-1.5

Francois

Re: [gentoo-science] [sage-on-gentoo] cannot build Sage because of Sympy-0.7.1

[v_2e]

On Thu, 16 Aug 2012 21:28:07 +1200
Francois Bissey <fbissey@slingshot.co.nz> wrote:

> you have numpy-1.6.xx installed don't you? I am adding that to the
> list of pains related to the fact we are stuck with numpy-1.5
  No, I don't. In fact, numpy-1.6 has been masked in
my /etc/portage/package.mask file for a while.
  Currently I've got dev-python/numpy-1.5.1 installed.

    Regards,
        Vladimir

----- 
 <v_2e@ukr.net>

Re: [gentoo-science] [sage-on-gentoo] cannot build Sage because of Sympy-0.7.1

[fbissey]

Quoting v_2e@ukr.net:

> On Thu, 16 Aug 2012 21:28:07 +1200
> Francois Bissey <fbissey@slingshot.co.nz> wrote:
>
>> you have numpy-1.6.xx installed don't you? I am adding that to the
>> list of pains related to the fact we are stuck with numpy-1.5
>   No, I don't. In fact, numpy-1.6 has been masked in
> my /etc/portage/package.mask file for a while.
>   Currently I've got dev-python/numpy-1.5.1 installed.
>
OK wrong assumption from me by looking at the code. You have probably enabled
the gmp useflag in mpmath and now you have gmpy on your system. I am 
now fairly
sure that gmpy is the cause of the problem.
Like sage gmpy can be autoloaded by mpmath but if both are present at the same
time  you get the confusion that you have seen. As a quick check you can try:
MPMATH_NOGMPY=1 emerge -1v sympy
that should work. Adding that variable to sage-env may help with potential
problems importing mpmath or sympy in sage but really mpmath has to change
their default behavior.

Francois

Re: [gentoo-science] [sage-on-gentoo] cannot build Sage because of Sympy-0.7.1

[v_2e]

  Hello!

On Fri, 17 Aug 2012 09:32:18 +1200
fbissey@slingshot.co.nz wrote:

> OK wrong assumption from me by looking at the code. You have probably
> enabled the gmp useflag in mpmath and now you have gmpy on your
> system. I am now fairly
> sure that gmpy is the cause of the problem.
> Like sage gmpy can be autoloaded by mpmath but if both are present at
> the same time  you get the confusion that you have seen. 
  Yes, the 'gmp' USE-flag is enabled for 'mpmath'.

> As a quick check you can try: MPMATH_NOGMPY=1 emerge -1v sympy
> that should work. Adding that variable to sage-env may help with
> potential problems importing mpmath or sympy in sage but really
> mpmath has to change their default behavior.
  So what do I have to do to have a working Sage? Do I have to rebuild
'mpmath' with the 'gmp' USE-flag disabled?

  Thanks!
    Vladimir

----- 
 <v_2e@ukr.net>

Re: [gentoo-science] [sage-on-gentoo] cannot build Sage because of Sympy-0.7.1

[Francois Bissey]

On Fri, 17 Aug 2012 12:20:12 v_2e@ukr.net wrote:
>   Hello!
> 
> On Fri, 17 Aug 2012 09:32:18 +1200
> 
> fbissey@slingshot.co.nz wrote:
> > OK wrong assumption from me by looking at the code. You have probably
> > enabled the gmp useflag in mpmath and now you have gmpy on your
> > system. I am now fairly
> > sure that gmpy is the cause of the problem.
> > Like sage gmpy can be autoloaded by mpmath but if both are present at
> > the same time  you get the confusion that you have seen.
> 
>   Yes, the 'gmp' USE-flag is enabled for 'mpmath'.
> 
> > As a quick check you can try: MPMATH_NOGMPY=1 emerge -1v sympy
> > that should work. Adding that variable to sage-env may help with
> > potential problems importing mpmath or sympy in sage but really
> > mpmath has to change their default behavior.
> 
>   So what do I have to do to have a working Sage? Do I have to rebuild
> 'mpmath' with the 'gmp' USE-flag disabled?
> 
Best solution: rebuild mpmath without the gmp useflag and remove gmpy.
Rebuilding is not enough as the flag only insure that gmpy is installed.
mpmath autodetect gmpy's presence all by itself if it is present - gmp flag or 
not.

If you don't want to do that you'll have to export MPMATH_NOGMPY=1
before using sage or emerging sympy or importing sympy or mpmath
from the python shell.

Francois

Re: [gentoo-science] [sage-on-gentoo] cannot build Sage because of Sympy-0.7.1

[v_2e]

On Fri, 17 Aug 2012 21:28:09 +1200
Francois Bissey <fbissey@slingshot.co.nz> wrote:

> Best solution: rebuild mpmath without the gmp useflag and remove gmpy.
> Rebuilding is not enough as the flag only insure that gmpy is
> installed. mpmath autodetect gmpy's presence all by itself if it is
> present - gmp flag or not.
> 
  I did as you suggested and now I can 'import mpmath' and 'import
sympy' in the Python2 shell without any errors.
  Maybe it is worth masking this 'gmp' USE-flag if Sage is being
emerged? Or is it impossible?

  Regards,
    Vladimir

----- 
 <v_2e@ukr.net>

Re: [gentoo-science] [sage-on-gentoo] cannot build Sage because of Sympy-0.7.1

[Francois Bissey]

On Fri, 17 Aug 2012 13:29:14 v_2e@ukr.net wrote:
> On Fri, 17 Aug 2012 21:28:09 +1200
> 
> Francois Bissey <fbissey@slingshot.co.nz> wrote:
> > Best solution: rebuild mpmath without the gmp useflag and remove gmpy.
> > Rebuilding is not enough as the flag only insure that gmpy is
> > installed. mpmath autodetect gmpy's presence all by itself if it is
> > present - gmp flag or not.
> 
>   I did as you suggested and now I can 'import mpmath' and 'import
> sympy' in the Python2 shell without any errors.
>   Maybe it is worth masking this 'gmp' USE-flag if Sage is being
> emerged? Or is it impossible?
> 
gmpy is the real culprit. You can have it independently of the useflag.
We can put something in the package.use we provide in the overlay.
We could ship a package.mask file as well.
I am setting MPMATH_NOGMPY in sage-env in sage-5.3 so there
won't be problems from sage.

Francois