Re: [sage-devel] Re: gcc problems on MacOS

[David Roe]

Thank you Matthias; clearing out /usr/local/include/flint* and
/usr/local/flint* solved the problem.
David

On Mon, Nov 20, 2023 at 5:51 PM Matthias Koeppe 
wrote:

> The real error from your log:
>
> sage/algebras/quatalg/quaternion_algebra_element.cpp:22735:3: error:
> 'fmpz_poly_scalar_mul_mpz' is deprecated. Use 'fmpz_poly_scalar_mul_fmpz'
> instead.
>   fmpz_poly_scalar_mul_mpz(__pyx_v_self->x, __pyx_v_self->x,
> __pyx_v_4sage_8algebras_7quatalg_26quaternion_algebra_element_T1);
>   ^
> ///usr/local/include/flint/fmpz_poly.h:1263:34: note: expanded from
> macro 'fmpz_poly_scalar_mul_mpz'
> #define fmpz_poly_scalar_mul_mpz _Pragma("GCC error
> \"'fmpz_poly_scalar_mul_mpz' is deprecated. Use 'fmpz_poly_scalar_mul_fmpz'
> instead.\"")
>  ^
> :433:6: note: expanded from here
>  GCC error "'fmpz_poly_scalar_mul_mpz' is deprecated. Use
> 'fmpz_poly_scalar_mul_fmpz' instead."
>  ^
>
> Your config.log says that you are using flint from SPKG. Above refers to a
> FLINT installation in /usr/local, likely incompatible FLINT 3. Best to
> uninstall that
>
>
> On Monday, November 20, 2023 at 2:46:18 PM UTC-8 David Roe wrote:
>
>> I'm trying to build Sage 10.2.rc4 on MacOS 13.3.1 and have been running
>> into multiple errors, some possibly due to Apple upgrading my OS when my
>> laptop went in for repair.  I've managed to get past all of Sage's
>> dependencies, and am currently stuck on an error in sagelib:
>>
>> g++ -std=gnu++11 -std=gnu++11 -bundle -undefined dynamic_lookup -isysroot
>> /Library/Developer/CommandLineTools/SDKs/MacOSX13.sdk
>> -L/Users/roed/sage/sage-10.2.rc4/local/lib -Wl,-rpath,
>>  -L/Users/roed/sage/sage-10.2.rc4/local/lib -lflint -lgmp -lSingular
>> -lpolys -lflint -lmpfr -lgmp -lfactory -lflint -lmpfr -lgmp -lntl -lgmp
>> -lomalloc -lsingular_resources -lgsl -lopenblas -lm -o
>> build/lib.macosx-13-x86_64-cpython-311/sage/symbolic/
>> expression.cpython-311-darwin.so -lpari
>> error: command '/Users/roed/sage/sage-10.2.rc4/local/libexec/ccache/gcc'
>> failed with exit code 1
>> (see config.log  and
>> sagelib-10.2.rc4.log  for
>> more details)
>>
>> Previously, I did
>> xcode-select --install
>> sudo xcode-select --switch /Library/Developer/CommandLineTools
>> brew install gcc
>> to get past a build error in scipy
>> 
>> ../meson.build:82:0: ERROR: Compiler gfortran cannot compile programs.
>>
>> I noticed that this is the first g++ call after a bunch of gcc calls, but
>> it doesn't seem to be the first g++ call overall.  Any ideas?
>> David
>>
> --
> You received this message because you are subscribed to the Google Groups
> "sage-devel" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to sage-devel+unsubscr...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/sage-devel/20f5c300-b82f-48f7-a0fa-a4e978f55030n%40googlegroups.com
> 
> .
>

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sage-devel/CAChs6_%3D2Oy0_kSG-%3DJxK%2BLnXN3yS6eCjE%3DKJ3%3DBMYB-tRHZmmg%40mail.gmail.com.

[sage-devel] Re: gcc problems on MacOS

[Matthias Koeppe]

The real error from your log:

sage/algebras/quatalg/quaternion_algebra_element.cpp:22735:3: error: 
'fmpz_poly_scalar_mul_mpz' is deprecated. Use 'fmpz_poly_scalar_mul_fmpz' 
instead.
  fmpz_poly_scalar_mul_mpz(__pyx_v_self->x, __pyx_v_self->x, 
__pyx_v_4sage_8algebras_7quatalg_26quaternion_algebra_element_T1);
  ^
///usr/local/include/flint/fmpz_poly.h:1263:34: note: expanded from 
macro 'fmpz_poly_scalar_mul_mpz'
#define fmpz_poly_scalar_mul_mpz _Pragma("GCC error 
\"'fmpz_poly_scalar_mul_mpz' is deprecated. Use 'fmpz_poly_scalar_mul_fmpz' 
instead.\"")
 ^
:433:6: note: expanded from here
 GCC error "'fmpz_poly_scalar_mul_mpz' is deprecated. Use 
'fmpz_poly_scalar_mul_fmpz' instead."
 ^

Your config.log says that you are using flint from SPKG. Above refers to a 
FLINT installation in /usr/local, likely incompatible FLINT 3. Best to 
uninstall that


On Monday, November 20, 2023 at 2:46:18 PM UTC-8 David Roe wrote:

> I'm trying to build Sage 10.2.rc4 on MacOS 13.3.1 and have been running 
> into multiple errors, some possibly due to Apple upgrading my OS when my 
> laptop went in for repair.  I've managed to get past all of Sage's 
> dependencies, and am currently stuck on an error in sagelib:
>
> g++ -std=gnu++11 -std=gnu++11 -bundle -undefined dynamic_lookup -isysroot 
> /Library/Developer/CommandLineTools/SDKs/MacOSX13.sdk 
> -L/Users/roed/sage/sage-10.2.rc4/local/lib -Wl,-rpath, 
>  -L/Users/roed/sage/sage-10.2.rc4/local/lib -lflint -lgmp -lSingular 
> -lpolys -lflint -lmpfr -lgmp -lfactory -lflint -lmpfr -lgmp -lntl -lgmp 
> -lomalloc -lsingular_resources -lgsl -lopenblas -lm -o 
> build/lib.macosx-13-x86_64-cpython-311/sage/symbolic/
> expression.cpython-311-darwin.so -lpari
> error: command '/Users/roed/sage/sage-10.2.rc4/local/libexec/ccache/gcc' 
> failed with exit code 1
> (see config.log  and 
> sagelib-10.2.rc4.log  for 
> more details)
>
> Previously, I did
> xcode-select --install
> sudo xcode-select --switch /Library/Developer/CommandLineTools
> brew install gcc
> to get past a build error in scipy 
> 
> ../meson.build:82:0: ERROR: Compiler gfortran cannot compile programs.
>
> I noticed that this is the first g++ call after a bunch of gcc calls, but 
> it doesn't seem to be the first g++ call overall.  Any ideas?
> David
>

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sage-devel/20f5c300-b82f-48f7-a0fa-a4e978f55030n%40googlegroups.com.

[sage-devel] Re: sage-10.2.rc4 does not build on opensuse tumbleweed

[Matthias Koeppe]

Our CI tests with the official opensuse/tumbleweed Docker image, 
see https://github.com/sagemath/sage/blob/develop/tox.ini#L381
I think it's updated daily.

On Monday, November 20, 2023 at 2:39:54 PM UTC-8 Christian Bär wrote:

> How recent is the tumbleweed install you are testing against? As 
> mentioned, sage-10.1 initially did compile until some tumbleweed update 
> broke it, leading to the error message described above. This is why I am 
> skeptical that installing more system packages will resolve the problem. 
>
> Matthias Koeppe schrieb am Sonntag, 19. November 2023 um 17:21:47 UTC+1:
>
>> We also see this failure on opensuse-tumbleweed in our CI in the 
>> "minimal" system package configuration 
>> https://github.com/sagemath/sage/actions/runs/6909825582/job/18801882455#step:11:2323
>>
>> However, the "standard" system package configuration builds and tests 
>> fine: 
>> https://github.com/sagemath/sage/actions/runs/6909825582/job/18806775962
>>
>> So likely this can be fixed on your system by installing more system 
>> packages.
>>
>> What's surprising in your config.log is that it is giving recommendations 
>> for using "apt-get" (excerpt below). Not sure where apt-get would be coming 
>> from, but if you move it away, then you'll receive system package 
>> recommendations for OpenSuSE (using zypper), as seen here: 
>> https://github.com/sagemath/sage/actions/runs/6909825582/job/18801882455#step:11:1722
>>
>> = config.log
>> notice: the following SPKGs did not find equivalent system packages:
>>
>> appdirs arb cachetools chardet colorama cypari distlib ecl eclib 
>> fflas_ffpack filelock flint fpylll freetype gengetopt giac givaro lcalc 
>> libgd linbox lrcalc mathjax matplotlib maxima meson nauty ncurses palp pari 
>> pari_galdata pari_seadata_small patchelf pillow pplpy primecount 
>> primecountpy primesieve python3 qhull readline rw singular sympow tox 
>> virtualenv   4ti2 _develop _recommended cbc coxeter3 gp2c igraph 
>> libgraphviz libnauty libxml2 lrslib pandoc pari_elldata pari_galpol 
>> pari_nftables pari_seadata pdf2svg perl_cpan_polymake_prereq perl_mongodb 
>> polymake
>> 
>> configure:154937: checking for the package system in use
>> configure:154940: result: debian
>> configure:154951:
>>
>> hint: installing the following system packages, if not
>> already present, is recommended and may avoid having to
>> build them (though some may have to be built anyway):
>>
>>   $ sudo apt-get update 
>>   $ sudo apt-get install  libflint-arb-dev ecl libec-dev eclib-tools 
>> fflas-ffpack libflint-dev libfreetype-dev gengetopt libgiac-dev xcas 
>> libgivaro-dev lcalc liblfunction-dev libgd-dev liblinbox-dev liblrcalc-dev 
>> maxima-sage maxima meson nauty libncurses5-dev palp pari-gp2c libpari-dev 
>> pari-doc pari-elldata pari-galdata pari-galpol pari-seadata patchelf 
>> libprimesieve-dev python3 libpython3-dev python3-distutils python3-venv 
>> libqhull-dev libreadline-dev librw-dev singular singular-doc 
>> libsingular4-dev sympow tox
>> =
>>
>>
>> On Sunday, November 19, 2023 at 8:03:07 AM UTC-8 Christian Bär wrote:
>>
>>> Dear all,
>>>
>>> sage-10.2.rc4 does not compile on opensuse tumbleweed (20231117). The 
>>> package arb does not build.  The package arb-devel is not installed on the 
>>> system because it is no longer available for tumbleweed. Sage compilation 
>>> stops with the error message:
>>>
>>> /bin/sh: symbol lookup error: /bin/sh: undefined symbol: 
>>> rl_trim_arg_from_keyseq, version READLINE_8.2
>>> make[8]: *** [../Makefile.subdirs:60: 
>>> ../build/fmpz_extras/lshift_mpn.lo] Error 127
>>> make[8]: *** [../Makefile.subdirs:60: 
>>> ../build/fmpz_extras/sub_small_large.lo] Error 127
>>> make[7]: *** [Makefile:128: libarb.so.2.14.0] Error 2
>>> make[6]: *** [Makefile:148: library] Error 2
>>> make[5]: *** [Makefile:95: verbose] Error 2
>>>
>>> 
>>> Error building arb-2.23.0
>>>
>>> 
>>>
>>> The log files are attached. The system has an AMD Ryzen 5950 CPU. 
>>>
>>> As a matter of fact, the same problem arises when I try to compile 
>>> sage-10.1 (with arb-2.22.1). When sage-10.1 came out, it did work 
>>> though. After some tumbleweed update 1 or 2 months ago, sage stopped 
>>> working and compilation was no longer possible.
>>>
>>> Best,
>>> Christian
>>>
>>

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sage-devel/a372aa16-c282-4ec9-85d9-32cd5dbe9e5cn%40googlegroups.com.

[sage-devel] gcc problems on MacOS

[David Roe]

I'm trying to build Sage 10.2.rc4 on MacOS 13.3.1 and have been running
into multiple errors, some possibly due to Apple upgrading my OS when my
laptop went in for repair.  I've managed to get past all of Sage's
dependencies, and am currently stuck on an error in sagelib:

g++ -std=gnu++11 -std=gnu++11 -bundle -undefined dynamic_lookup -isysroot
/Library/Developer/CommandLineTools/SDKs/MacOSX13.sdk
-L/Users/roed/sage/sage-10.2.rc4/local/lib -Wl,-rpath,
 -L/Users/roed/sage/sage-10.2.rc4/local/lib -lflint -lgmp -lSingular
-lpolys -lflint -lmpfr -lgmp -lfactory -lflint -lmpfr -lgmp -lntl -lgmp
-lomalloc -lsingular_resources -lgsl -lopenblas -lm -o
build/lib.macosx-13-x86_64-cpython-311/sage/symbolic/
expression.cpython-311-darwin.so -lpari
error: command '/Users/roed/sage/sage-10.2.rc4/local/libexec/ccache/gcc'
failed with exit code 1
(see config.log  and
sagelib-10.2.rc4.log  for
more details)

Previously, I did
xcode-select --install
sudo xcode-select --switch /Library/Developer/CommandLineTools
brew install gcc
to get past a build error in scipy

../meson.build:82:0: ERROR: Compiler gfortran cannot compile programs.

I noticed that this is the first g++ call after a bunch of gcc calls, but
it doesn't seem to be the first g++ call overall.  Any ideas?
David

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sage-devel/CAChs6_mYLxHuyCqpFTs09nmr0Z9J2fYeBuMA4pvh3a1j_PVq5w%40mail.gmail.com.

[sage-devel] Re: sage-10.2.rc4 does not build on opensuse tumbleweed

[Christian Bär]

How recent is the tumbleweed install you are testing against? As mentioned, 
sage-10.1 initially did compile until some tumbleweed update broke it, 
leading to the error message described above. This is why I am skeptical 
that installing more system packages will resolve the problem. 

Matthias Koeppe schrieb am Sonntag, 19. November 2023 um 17:21:47 UTC+1:

> We also see this failure on opensuse-tumbleweed in our CI in the "minimal" 
> system package configuration 
> https://github.com/sagemath/sage/actions/runs/6909825582/job/18801882455#step:11:2323
>
> However, the "standard" system package configuration builds and tests 
> fine: 
> https://github.com/sagemath/sage/actions/runs/6909825582/job/18806775962
>
> So likely this can be fixed on your system by installing more system 
> packages.
>
> What's surprising in your config.log is that it is giving recommendations 
> for using "apt-get" (excerpt below). Not sure where apt-get would be coming 
> from, but if you move it away, then you'll receive system package 
> recommendations for OpenSuSE (using zypper), as seen here: 
> https://github.com/sagemath/sage/actions/runs/6909825582/job/18801882455#step:11:1722
>
> = config.log
> notice: the following SPKGs did not find equivalent system packages:
>
> appdirs arb cachetools chardet colorama cypari distlib ecl eclib 
> fflas_ffpack filelock flint fpylll freetype gengetopt giac givaro lcalc 
> libgd linbox lrcalc mathjax matplotlib maxima meson nauty ncurses palp pari 
> pari_galdata pari_seadata_small patchelf pillow pplpy primecount 
> primecountpy primesieve python3 qhull readline rw singular sympow tox 
> virtualenv   4ti2 _develop _recommended cbc coxeter3 gp2c igraph 
> libgraphviz libnauty libxml2 lrslib pandoc pari_elldata pari_galpol 
> pari_nftables pari_seadata pdf2svg perl_cpan_polymake_prereq perl_mongodb 
> polymake
> 
> configure:154937: checking for the package system in use
> configure:154940: result: debian
> configure:154951:
>
> hint: installing the following system packages, if not
> already present, is recommended and may avoid having to
> build them (though some may have to be built anyway):
>
>   $ sudo apt-get update 
>   $ sudo apt-get install  libflint-arb-dev ecl libec-dev eclib-tools 
> fflas-ffpack libflint-dev libfreetype-dev gengetopt libgiac-dev xcas 
> libgivaro-dev lcalc liblfunction-dev libgd-dev liblinbox-dev liblrcalc-dev 
> maxima-sage maxima meson nauty libncurses5-dev palp pari-gp2c libpari-dev 
> pari-doc pari-elldata pari-galdata pari-galpol pari-seadata patchelf 
> libprimesieve-dev python3 libpython3-dev python3-distutils python3-venv 
> libqhull-dev libreadline-dev librw-dev singular singular-doc 
> libsingular4-dev sympow tox
> =
>
>
> On Sunday, November 19, 2023 at 8:03:07 AM UTC-8 Christian Bär wrote:
>
>> Dear all,
>>
>> sage-10.2.rc4 does not compile on opensuse tumbleweed (20231117). The 
>> package arb does not build.  The package arb-devel is not installed on the 
>> system because it is no longer available for tumbleweed. Sage compilation 
>> stops with the error message:
>>
>> /bin/sh: symbol lookup error: /bin/sh: undefined symbol: 
>> rl_trim_arg_from_keyseq, version READLINE_8.2
>> make[8]: *** [../Makefile.subdirs:60: ../build/fmpz_extras/lshift_mpn.lo] 
>> Error 127
>> make[8]: *** [../Makefile.subdirs:60: 
>> ../build/fmpz_extras/sub_small_large.lo] Error 127
>> make[7]: *** [Makefile:128: libarb.so.2.14.0] Error 2
>> make[6]: *** [Makefile:148: library] Error 2
>> make[5]: *** [Makefile:95: verbose] Error 2
>>
>> 
>> Error building arb-2.23.0
>>
>> 
>>
>> The log files are attached. The system has an AMD Ryzen 5950 CPU. 
>>
>> As a matter of fact, the same problem arises when I try to compile 
>> sage-10.1 (with arb-2.22.1). When sage-10.1 came out, it did work 
>> though. After some tumbleweed update 1 or 2 months ago, sage stopped 
>> working and compilation was no longer possible.
>>
>> Best,
>> Christian
>>
>

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sage-devel/ff29e330-40c8-43d7-9453-a9d93ed188a3n%40googlegroups.com.