Re: [sage-devel] Sage 9.4.beta3 failed to build (gcc-10.3.0)

2021-06-22 Thread Kazuyoshi Furutaka 
😳

2021年6月23日(水) 7:57 Volker Braun :

> I'm running Fedora 34 and 9.4.beta3 builds fine (x86_64). It did compile
> the included gcc.
>
>
>
>
> On Tuesday, June 22, 2021 at 11:50:00 PM UTC+2 dim...@gmail.com wrote:
>
>>
>>
>> On Tue, 22 Jun 2021, 22:16 Kazuyoshi Furutaka, 
>> wrote:
>>
>>> On the other hand, build of gcc-10.3 failed.   Then, what to do?
>>>
>>
>> use another compiler.
>> e.g. you can use clang
>>
>> dnf install clang
>>
>> then run Sages configure as follows
>>
>> CC=clang CXX=clang++ ./configure
>>
>> and then
>>
>> make
>>
>>
>> This configuration is not tested much on Linux, but on macOS this is the
>> compiler used in our builds.
>>
>>
>>
>>> Kazuyoshi
>>>
>>> 2021年6月23日(水) 5:40 Dima Pasechnik :
>>>
 Sage is not yet ready to be built with gcc 11, see
 https://trac.sagemath.org/ticket/31786

 On Tue, Jun 22, 2021 at 8:36 PM Kazuyoshi Furutaka
  wrote:
 >
 >
 >
 > 2021年6月23日(水) 0:06 Dima Pasechnik :
 >>
 >> On Tue, Jun 22, 2021 at 3:45 PM Kazuyoshi Furutaka
 >>  wrote:
 >> [...]
 >> > First of all, I'd like to let you know that on Fedora-34 (x86_64),
 the simple build (configured with --prefix only) of prestine gcc-10.3.0
 fails without the static version of libstdc++ library (please read this
 thread for details).
 >> >
 >> > On Fedora-33 systems, it can build without libstdc++.a; I've tried
 to investigate the reason but couldn't figure out yet.
 >> >
 >> > Then, even on a Fedora-34 (x86_64) system WITH libstdc++.a, the
 build of Sage 9.4.beta3 fails; the build of gcc-10.3.0 failed due to
 "gcc-multilib-multiarch.patch": without it, the build of gcc succeeds.
 I'll attach the log of the failed gcc build.
 >> > On Fedora,
 >> >   MULTILIB_OSDIRNAMES =  m64=../lib64 m32=../lib mx32=../libx32
 >> > instead of
 >> >   MULTILIB_OSDIRNAMES =  m64=../lib m32=../lib32 mx32=../libx32
 >> >
 >> > Sorry for just a report; I could not yet figure out how to resolve
 this issue (modify the patch).
 >>
 >> it is not clear what you are trying to do.
 >>
 >> The reasons gcc is included as a package in Sage are mostly
 >> historical; I don't think there are many current use cases for it,
 >> and I think it actually should be removed.
 >> Fortunately there are enough options nowadays to source C/C++/Fortran
 >> compilers elsewhere.
 >> (at most, Sage might need to add some C,C++,Fortran flags)
 >>
 >> HTH,
 >> Dima
 >
 >
 > I once post a reply but it seems to disappear...
 >
 > I just want to build Sage from source.
 > The sage-build try to build GCC and a huge number of its dependent
 packages though they're already on my system as rpm packages because GCC on
 Fedora is TOO NEW.
 > So I'm very happy if someone modify the build system to avoid the
 time-consuming build of GCC and its dependent!
 >
 > Yours,
 > Kazuyoshi
 > --
 > Kazuyoshi Furutaka
 >
 > --
 > 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+...@googlegroups.com.
 > To view this discussion on the web visit
 https://groups.google.com/d/msgid/sage-devel/CANjERvf_M%3DDf25S3HKvkPhzXMxNEXGRcAbtUSU6g0J2UG6Q3CQ%40mail.gmail.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+...@googlegroups.com.
 To view this discussion on the web visit
 https://groups.google.com/d/msgid/sage-devel/CAAWYfq2g68RNyD6L6jw2d_XNtOrdrqc8K3TvoMcU_H9iVXRPLQ%40mail.gmail.com
 .

>>>
>>>
>>> --
>>> Kazuyoshi Furutaka
>>>
>>> --
>>> 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+...@googlegroups.com.
>>>
>> To view this discussion on the web visit
>>> https://groups.google.com/d/msgid/sage-devel/CANjERveKpCrg%2BhJ-Expdkx5is-Jy7ehhhmmw3buWzcBpPL9o7w%40mail.gmail.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/a2dc0758-8dd3-4b59-aa28-27d9d2f5e461n%40googlegroups.com
>

Re: [sage-devel] Sage 9.4.beta3 failed to build (gcc-10.3.0)

2021-06-22 Thread Volker Braun 
I'm running Fedora 34 and 9.4.beta3 builds fine (x86_64). It did compile 
the included gcc.




On Tuesday, June 22, 2021 at 11:50:00 PM UTC+2 dim...@gmail.com wrote:

>
>
> On Tue, 22 Jun 2021, 22:16 Kazuyoshi Furutaka,  
> wrote:
>
>> On the other hand, build of gcc-10.3 failed.   Then, what to do?
>>
>
> use another compiler.
> e.g. you can use clang
>
> dnf install clang
>
> then run Sages configure as follows
>
> CC=clang CXX=clang++ ./configure
>
> and then
>
> make
>
>
> This configuration is not tested much on Linux, but on macOS this is the 
> compiler used in our builds.
>
>
>
>> Kazuyoshi
>>
>> 2021年6月23日(水) 5:40 Dima Pasechnik :
>>
>>> Sage is not yet ready to be built with gcc 11, see
>>> https://trac.sagemath.org/ticket/31786
>>>
>>> On Tue, Jun 22, 2021 at 8:36 PM Kazuyoshi Furutaka
>>>  wrote:
>>> >
>>> >
>>> >
>>> > 2021年6月23日(水) 0:06 Dima Pasechnik :
>>> >>
>>> >> On Tue, Jun 22, 2021 at 3:45 PM Kazuyoshi Furutaka
>>> >>  wrote:
>>> >> [...]
>>> >> > First of all, I'd like to let you know that on Fedora-34 (x86_64), 
>>> the simple build (configured with --prefix only) of prestine gcc-10.3.0 
>>> fails without the static version of libstdc++ library (please read this 
>>> thread for details).
>>> >> >
>>> >> > On Fedora-33 systems, it can build without libstdc++.a; I've tried 
>>> to investigate the reason but couldn't figure out yet.
>>> >> >
>>> >> > Then, even on a Fedora-34 (x86_64) system WITH libstdc++.a, the 
>>> build of Sage 9.4.beta3 fails; the build of gcc-10.3.0 failed due to 
>>> "gcc-multilib-multiarch.patch": without it, the build of gcc succeeds.  
>>> I'll attach the log of the failed gcc build.
>>> >> > On Fedora,
>>> >> >   MULTILIB_OSDIRNAMES =  m64=../lib64 m32=../lib mx32=../libx32
>>> >> > instead of
>>> >> >   MULTILIB_OSDIRNAMES =  m64=../lib m32=../lib32 mx32=../libx32
>>> >> >
>>> >> > Sorry for just a report; I could not yet figure out how to resolve 
>>> this issue (modify the patch).
>>> >>
>>> >> it is not clear what you are trying to do.
>>> >>
>>> >> The reasons gcc is included as a package in Sage are mostly
>>> >> historical; I don't think there are many current use cases for it,
>>> >> and I think it actually should be removed.
>>> >> Fortunately there are enough options nowadays to source C/C++/Fortran
>>> >> compilers elsewhere.
>>> >> (at most, Sage might need to add some C,C++,Fortran flags)
>>> >>
>>> >> HTH,
>>> >> Dima
>>> >
>>> >
>>> > I once post a reply but it seems to disappear...
>>> >
>>> > I just want to build Sage from source.
>>> > The sage-build try to build GCC and a huge number of its dependent 
>>> packages though they're already on my system as rpm packages because GCC on 
>>> Fedora is TOO NEW.
>>> > So I'm very happy if someone modify the build system to avoid the 
>>> time-consuming build of GCC and its dependent!
>>> >
>>> > Yours,
>>> > Kazuyoshi
>>> > --
>>> > Kazuyoshi Furutaka
>>> >
>>> > --
>>> > 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+...@googlegroups.com.
>>> > To view this discussion on the web visit 
>>> https://groups.google.com/d/msgid/sage-devel/CANjERvf_M%3DDf25S3HKvkPhzXMxNEXGRcAbtUSU6g0J2UG6Q3CQ%40mail.gmail.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+...@googlegroups.com.
>>> To view this discussion on the web visit 
>>> https://groups.google.com/d/msgid/sage-devel/CAAWYfq2g68RNyD6L6jw2d_XNtOrdrqc8K3TvoMcU_H9iVXRPLQ%40mail.gmail.com
>>> .
>>>
>>
>>
>> -- 
>> Kazuyoshi Furutaka
>>
>> -- 
>> 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+...@googlegroups.com.
>>
> To view this discussion on the web visit 
>> https://groups.google.com/d/msgid/sage-devel/CANjERveKpCrg%2BhJ-Expdkx5is-Jy7ehhhmmw3buWzcBpPL9o7w%40mail.gmail.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/a2dc0758-8dd3-4b59-aa28-27d9d2f5e461n%40googlegroups.com.

Re: [sage-devel] Sage 9.4.beta3 failed to build (gcc-10.3.0)

2021-06-22 Thread Dima Pasechnik 
On Tue, 22 Jun 2021, 22:16 Kazuyoshi Furutaka, 
wrote:

> On the other hand, build of gcc-10.3 failed.   Then, what to do?
>

use another compiler.
e.g. you can use clang

dnf install clang

then run Sages configure as follows

CC=clang CXX=clang++ ./configure

and then

make


This configuration is not tested much on Linux, but on macOS this is the
compiler used in our builds.



> Kazuyoshi
>
> 2021年6月23日(水) 5:40 Dima Pasechnik :
>
>> Sage is not yet ready to be built with gcc 11, see
>> https://trac.sagemath.org/ticket/31786
>>
>> On Tue, Jun 22, 2021 at 8:36 PM Kazuyoshi Furutaka
>>  wrote:
>> >
>> >
>> >
>> > 2021年6月23日(水) 0:06 Dima Pasechnik :
>> >>
>> >> On Tue, Jun 22, 2021 at 3:45 PM Kazuyoshi Furutaka
>> >>  wrote:
>> >> [...]
>> >> > First of all, I'd like to let you know that on Fedora-34 (x86_64),
>> the simple build (configured with --prefix only) of prestine gcc-10.3.0
>> fails without the static version of libstdc++ library (please read this
>> thread for details).
>> >> >
>> >> > On Fedora-33 systems, it can build without libstdc++.a; I've tried
>> to investigate the reason but couldn't figure out yet.
>> >> >
>> >> > Then, even on a Fedora-34 (x86_64) system WITH libstdc++.a, the
>> build of Sage 9.4.beta3 fails; the build of gcc-10.3.0 failed due to
>> "gcc-multilib-multiarch.patch": without it, the build of gcc succeeds.
>> I'll attach the log of the failed gcc build.
>> >> > On Fedora,
>> >> >   MULTILIB_OSDIRNAMES =  m64=../lib64 m32=../lib mx32=../libx32
>> >> > instead of
>> >> >   MULTILIB_OSDIRNAMES =  m64=../lib m32=../lib32 mx32=../libx32
>> >> >
>> >> > Sorry for just a report; I could not yet figure out how to resolve
>> this issue (modify the patch).
>> >>
>> >> it is not clear what you are trying to do.
>> >>
>> >> The reasons gcc is included as a package in Sage are mostly
>> >> historical; I don't think there are many current use cases for it,
>> >> and I think it actually should be removed.
>> >> Fortunately there are enough options nowadays to source C/C++/Fortran
>> >> compilers elsewhere.
>> >> (at most, Sage might need to add some C,C++,Fortran flags)
>> >>
>> >> HTH,
>> >> Dima
>> >
>> >
>> > I once post a reply but it seems to disappear...
>> >
>> > I just want to build Sage from source.
>> > The sage-build try to build GCC and a huge number of its dependent
>> packages though they're already on my system as rpm packages because GCC on
>> Fedora is TOO NEW.
>> > So I'm very happy if someone modify the build system to avoid the
>> time-consuming build of GCC and its dependent!
>> >
>> > Yours,
>> > Kazuyoshi
>> > --
>> > Kazuyoshi Furutaka
>> >
>> > --
>> > 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/CANjERvf_M%3DDf25S3HKvkPhzXMxNEXGRcAbtUSU6g0J2UG6Q3CQ%40mail.gmail.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/CAAWYfq2g68RNyD6L6jw2d_XNtOrdrqc8K3TvoMcU_H9iVXRPLQ%40mail.gmail.com
>> .
>>
>
>
> --
> Kazuyoshi Furutaka
>
> --
> 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/CANjERveKpCrg%2BhJ-Expdkx5is-Jy7ehhhmmw3buWzcBpPL9o7w%40mail.gmail.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/CAAWYfq04iN8oMAtugh%3DZmfSCoZ54zg%3DNPQ8qNmK2Bqejghy60w%40mail.gmail.com.

Re: [sage-devel] Sage 9.4.beta3 failed to build (gcc-10.3.0)

2021-06-22 Thread Kazuyoshi Furutaka 
On the other hand, build of gcc-10.3 failed.   Then, what to do?

Kazuyoshi

2021年6月23日(水) 5:40 Dima Pasechnik :

> Sage is not yet ready to be built with gcc 11, see
> https://trac.sagemath.org/ticket/31786
>
> On Tue, Jun 22, 2021 at 8:36 PM Kazuyoshi Furutaka
>  wrote:
> >
> >
> >
> > 2021年6月23日(水) 0:06 Dima Pasechnik :
> >>
> >> On Tue, Jun 22, 2021 at 3:45 PM Kazuyoshi Furutaka
> >>  wrote:
> >> [...]
> >> > First of all, I'd like to let you know that on Fedora-34 (x86_64),
> the simple build (configured with --prefix only) of prestine gcc-10.3.0
> fails without the static version of libstdc++ library (please read this
> thread for details).
> >> >
> >> > On Fedora-33 systems, it can build without libstdc++.a; I've tried to
> investigate the reason but couldn't figure out yet.
> >> >
> >> > Then, even on a Fedora-34 (x86_64) system WITH libstdc++.a, the build
> of Sage 9.4.beta3 fails; the build of gcc-10.3.0 failed due to
> "gcc-multilib-multiarch.patch": without it, the build of gcc succeeds.
> I'll attach the log of the failed gcc build.
> >> > On Fedora,
> >> >   MULTILIB_OSDIRNAMES =  m64=../lib64 m32=../lib mx32=../libx32
> >> > instead of
> >> >   MULTILIB_OSDIRNAMES =  m64=../lib m32=../lib32 mx32=../libx32
> >> >
> >> > Sorry for just a report; I could not yet figure out how to resolve
> this issue (modify the patch).
> >>
> >> it is not clear what you are trying to do.
> >>
> >> The reasons gcc is included as a package in Sage are mostly
> >> historical; I don't think there are many current use cases for it,
> >> and I think it actually should be removed.
> >> Fortunately there are enough options nowadays to source C/C++/Fortran
> >> compilers elsewhere.
> >> (at most, Sage might need to add some C,C++,Fortran flags)
> >>
> >> HTH,
> >> Dima
> >
> >
> > I once post a reply but it seems to disappear...
> >
> > I just want to build Sage from source.
> > The sage-build try to build GCC and a huge number of its dependent
> packages though they're already on my system as rpm packages because GCC on
> Fedora is TOO NEW.
> > So I'm very happy if someone modify the build system to avoid the
> time-consuming build of GCC and its dependent!
> >
> > Yours,
> > Kazuyoshi
> > --
> > Kazuyoshi Furutaka
> >
> > --
> > 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/CANjERvf_M%3DDf25S3HKvkPhzXMxNEXGRcAbtUSU6g0J2UG6Q3CQ%40mail.gmail.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/CAAWYfq2g68RNyD6L6jw2d_XNtOrdrqc8K3TvoMcU_H9iVXRPLQ%40mail.gmail.com
> .
>


-- 
Kazuyoshi Furutaka

-- 
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/CANjERveKpCrg%2BhJ-Expdkx5is-Jy7ehhhmmw3buWzcBpPL9o7w%40mail.gmail.com.

Re: [sage-devel] Sage 9.4.beta3 failed to build (gcc-10.3.0)

2021-06-22 Thread Dima Pasechnik 
Sage is not yet ready to be built with gcc 11, see
https://trac.sagemath.org/ticket/31786

On Tue, Jun 22, 2021 at 8:36 PM Kazuyoshi Furutaka
 wrote:
>
>
>
> 2021年6月23日(水) 0:06 Dima Pasechnik :
>>
>> On Tue, Jun 22, 2021 at 3:45 PM Kazuyoshi Furutaka
>>  wrote:
>> [...]
>> > First of all, I'd like to let you know that on Fedora-34 (x86_64), the 
>> > simple build (configured with --prefix only) of prestine gcc-10.3.0 fails 
>> > without the static version of libstdc++ library (please read this thread 
>> > for details).
>> >
>> > On Fedora-33 systems, it can build without libstdc++.a; I've tried to 
>> > investigate the reason but couldn't figure out yet.
>> >
>> > Then, even on a Fedora-34 (x86_64) system WITH libstdc++.a, the build of 
>> > Sage 9.4.beta3 fails; the build of gcc-10.3.0 failed due to 
>> > "gcc-multilib-multiarch.patch": without it, the build of gcc succeeds.  
>> > I'll attach the log of the failed gcc build.
>> > On Fedora,
>> >   MULTILIB_OSDIRNAMES =  m64=../lib64 m32=../lib mx32=../libx32
>> > instead of
>> >   MULTILIB_OSDIRNAMES =  m64=../lib m32=../lib32 mx32=../libx32
>> >
>> > Sorry for just a report; I could not yet figure out how to resolve this 
>> > issue (modify the patch).
>>
>> it is not clear what you are trying to do.
>>
>> The reasons gcc is included as a package in Sage are mostly
>> historical; I don't think there are many current use cases for it,
>> and I think it actually should be removed.
>> Fortunately there are enough options nowadays to source C/C++/Fortran
>> compilers elsewhere.
>> (at most, Sage might need to add some C,C++,Fortran flags)
>>
>> HTH,
>> Dima
>
>
> I once post a reply but it seems to disappear...
>
> I just want to build Sage from source.
> The sage-build try to build GCC and a huge number of its dependent packages 
> though they're already on my system as rpm packages because GCC on Fedora is 
> TOO NEW.
> So I'm very happy if someone modify the build system to avoid the 
> time-consuming build of GCC and its dependent!
>
> Yours,
> Kazuyoshi
> --
> Kazuyoshi Furutaka
>
> --
> 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/CANjERvf_M%3DDf25S3HKvkPhzXMxNEXGRcAbtUSU6g0J2UG6Q3CQ%40mail.gmail.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/CAAWYfq2g68RNyD6L6jw2d_XNtOrdrqc8K3TvoMcU_H9iVXRPLQ%40mail.gmail.com.

Re: [sage-devel] Sage 9.4.beta3 failed to build (gcc-10.3.0)

2021-06-22 Thread Kazuyoshi Furutaka 


2021年6月23日水曜日 0:06:44 UTC+9 dim...@gmail.com:

> On Tue, Jun 22, 2021 at 3:45 PM Kazuyoshi Furutaka 
>  wrote: 
> [...] 
> > First of all, I'd like to let you know that on Fedora-34 (x86_64), the 
> simple build (configured with --prefix only) of prestine gcc-10.3.0 fails 
> without the static version of libstdc++ library (please read this thread 
> for details). 
> > 
> > On Fedora-33 systems, it can build without libstdc++.a; I've tried to 
> investigate the reason but couldn't figure out yet. 
> > 
> > Then, even on a Fedora-34 (x86_64) system WITH libstdc++.a, the build of 
> Sage 9.4.beta3 fails; the build of gcc-10.3.0 failed due to 
> "gcc-multilib-multiarch.patch": without it, the build of gcc succeeds. I'll 
> attach the log of the failed gcc build. 
> > On Fedora, 
> > MULTILIB_OSDIRNAMES = m64=../lib64 m32=../lib mx32=../libx32 
> > instead of 
> > MULTILIB_OSDIRNAMES = m64=../lib m32=../lib32 mx32=../libx32 
> > 
> > Sorry for just a report; I could not yet figure out how to resolve this 
> issue (modify the patch). 
>
> it is not clear what you are trying to do. 
>
> The reasons gcc is included as a package in Sage are mostly 
> historical; I don't think there are many current use cases for it, 
> and I think it actually should be removed. 
> Fortunately there are enough options nowadays to source C/C++/Fortran 
> compilers elsewhere. 
> (at most, Sage might need to add some C,C++,Fortran flags) 
>
> HTH, 
> Dima
>

I just do want to merely build Sage from source, but the build failed on 
current Fedora linux due to GCC build (version 11 already exits on 
Fedora-34).
I do NOT want to do time-consuming compile of GCC and the SPKGs dependent 
on it for Sage because most of them are ALREADY on our systems as rpm 
packages, but Sage try to do so but only to fail to do one of them.
At first, I tried to figure out why the Sage-build can NOT find the GCC on 
my Fedora; it is because my GCC is too NEW and so the Sage-build is 
compiling the older GCC and a huge number of its dependent packages.
So I'm very happy if someone modify the Sage-build so that Sage does NOT 
try to build GCC and its dependent (if its on a system) ASAP.

Yours,
Kazuyoshi

-- 
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/48edf74e-cb67-4ecc-9a9b-ef7d57be644en%40googlegroups.com.

Re: [sage-devel] Sage 9.4.beta3 failed to build (gcc-10.3.0)

2021-06-22 Thread Kazuyoshi Furutaka 
2021年6月23日(水) 0:06 Dima Pasechnik :

> On Tue, Jun 22, 2021 at 3:45 PM Kazuyoshi Furutaka
>  wrote:
> [...]
> > First of all, I'd like to let you know that on Fedora-34 (x86_64), the
> simple build (configured with --prefix only) of prestine gcc-10.3.0 fails
> without the static version of libstdc++ library (please read this thread
> for details).
> >
> > On Fedora-33 systems, it can build without libstdc++.a; I've tried to
> investigate the reason but couldn't figure out yet.
> >
> > Then, even on a Fedora-34 (x86_64) system WITH libstdc++.a, the build of
> Sage 9.4.beta3 fails; the build of gcc-10.3.0 failed due to
> "gcc-multilib-multiarch.patch": without it, the build of gcc succeeds.
> I'll attach the log of the failed gcc build.
> > On Fedora,
> >   MULTILIB_OSDIRNAMES =  m64=../lib64 m32=../lib mx32=../libx32
> > instead of
> >   MULTILIB_OSDIRNAMES =  m64=../lib m32=../lib32 mx32=../libx32
> >
> > Sorry for just a report; I could not yet figure out how to resolve this
> issue (modify the patch).
>
> it is not clear what you are trying to do.
>
> The reasons gcc is included as a package in Sage are mostly
> historical; I don't think there are many current use cases for it,
> and I think it actually should be removed.
> Fortunately there are enough options nowadays to source C/C++/Fortran
> compilers elsewhere.
> (at most, Sage might need to add some C,C++,Fortran flags)
>
> HTH,
> Dima
>

I once post a reply but it seems to disappear...

I just want to build Sage from source.
The sage-build try to build GCC and a huge number of its dependent packages
though they're already on my system as rpm packages because GCC on Fedora
is TOO NEW.
So I'm very happy if someone modify the build system to avoid the
time-consuming build of GCC and its dependent!

Yours,
Kazuyoshi
-- 
Kazuyoshi Furutaka

-- 
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/CANjERvf_M%3DDf25S3HKvkPhzXMxNEXGRcAbtUSU6g0J2UG6Q3CQ%40mail.gmail.com.

Re: [sage-devel] Error compiling Cython file

2021-06-22 Thread Dima Pasechnik 
sagelib-9.4.beta1 error is just an uncleaned log file.

Do you build Python3 and not use one from the system?
(this would have avoided a potential problem building python)

You can post details of errors from logs/install.log, and your changes as a
diff or a link to
git branch somewhere (say, on github) -  otherwise it's hard to tell what's
wrong.


On Tue, 22 Jun 2021, 18:13 Hongli Zhao,  wrote:

> Hi all,
>
> I'm attempting to rebuild sage by running ```make``` in order to test some
> edits to the source code (```./sage``` was working before).
>
> However, ```make``` shows that 3 packages failed to build:
>
>
> Error building Sage.
>
>
> The following package(s) may have failed to build (not necessarily
>
> during this run of 'make all-start'):
> * package: sagelib-9.4.beta1
>
> * package: sagelib-9.4.beta3
>
> * package: python3-3.9.5
>
>
> When I checked the individual .log files I found that many functions
> seemed to have failed due to:
>
> "Error compiling Cython file"
>
>
> Could someone help me identify what the problem might be? Please let me
> know if my description is too vague. Thanks!
>
>
> --
> 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/e4ebab98-b074-419d-a7a6-50d2885a777an%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/CAAWYfq2NRY-%3DzEFTeXH1bo4c6d%2BGSrvj6VdsLbXZkpUSpLDnFA%40mail.gmail.com.

[sage-devel] Error compiling Cython file

2021-06-22 Thread Hongli Zhao 
Hi all,

I'm attempting to rebuild sage by running ```make``` in order to test some 
edits to the source code (```./sage``` was working before). 

However, ```make``` shows that 3 packages failed to build:


Error building Sage.


The following package(s) may have failed to build (not necessarily

during this run of 'make all-start'):
* package: sagelib-9.4.beta1

* package: sagelib-9.4.beta3

* package: python3-3.9.5


When I checked the individual .log files I found that many functions seemed 
to have failed due to:

"Error compiling Cython file"


Could someone help me identify what the problem might be? Please let me 
know if my description is too vague. Thanks!


-- 
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/e4ebab98-b074-419d-a7a6-50d2885a777an%40googlegroups.com.

[sage-devel] lcm broken for some polynomial rings

2021-06-22 Thread atg...@gmail.com 
Hello all,

The following code currently fails:

sage: K. = GF(2)[]
sage: lcm(K(0), K(0))

In polynomial element, the lcm code attempts to divide by the gcd of the 
two polynomials, throwing a zero division error in this case. One possible 
fix would be to add a try except block to lcm in polynomial element to 
catch this error.

I opened ticket 32033  for this 
error.

Can anyone more knowledgeable about the polynomial implementation comment 
on this?

Thank you,
Alex

-- 
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/9517fa8f-2fa8-4a46-9cef-943c8850957en%40googlegroups.com.

Re: [sage-devel] Sage 9.4.beta3 failed to build (gcc-10.3.0)

2021-06-22 Thread Dima Pasechnik 
On Tue, Jun 22, 2021 at 3:45 PM Kazuyoshi Furutaka
 wrote:
[...]
> First of all, I'd like to let you know that on Fedora-34 (x86_64), the simple 
> build (configured with --prefix only) of prestine gcc-10.3.0 fails without 
> the static version of libstdc++ library (please read this thread for details).
>
> On Fedora-33 systems, it can build without libstdc++.a; I've tried to 
> investigate the reason but couldn't figure out yet.
>
> Then, even on a Fedora-34 (x86_64) system WITH libstdc++.a, the build of Sage 
> 9.4.beta3 fails; the build of gcc-10.3.0 failed due to 
> "gcc-multilib-multiarch.patch": without it, the build of gcc succeeds.  I'll 
> attach the log of the failed gcc build.
> On Fedora,
>   MULTILIB_OSDIRNAMES =  m64=../lib64 m32=../lib mx32=../libx32
> instead of
>   MULTILIB_OSDIRNAMES =  m64=../lib m32=../lib32 mx32=../libx32
>
> Sorry for just a report; I could not yet figure out how to resolve this issue 
> (modify the patch).

it is not clear what you are trying to do.

The reasons gcc is included as a package in Sage are mostly
historical; I don't think there are many current use cases for it,
and I think it actually should be removed.
Fortunately there are enough options nowadays to source C/C++/Fortran
compilers elsewhere.
(at most, Sage might need to add some C,C++,Fortran flags)

HTH,
Dima
>
> Yours,
> Kazuyoshi
>
> --
> 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/787484d6-5ab9-4fdd-b909-2058bbbee5den%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/CAAWYfq3ggNg7YgDwesNy-926MW141H9WirnCjab6u4qQy2BPcw%40mail.gmail.com.

[sage-devel] downloading of packages from mirrors hangs

2021-06-22 Thread Dima Pasechnik 
Then, upon hitting ctrl-c I see
something like

[fplll-5.4.1] app.download_tarball(args.url_or_tarball,
args.destination, args.allow_upstream)
[fplll-5.4.1]   File
"/home/dima/sage/dev/sage/build/bin/../sage_bootstrap/download/app.py",
line 41, in download_tarball
[fplll-5.4.1] tarball.download(allow_upstream=allow_upstream)
[fplll-5.4.1]   File
"/home/dima/sage/dev/sage/build/bin/../sage_bootstrap/tarball.py",
line 161, in download
[fplll-5.4.1] for mirror in MirrorList():
[fplll-5.4.1]   File
"/home/dima/sage/dev/sage/build/bin/../sage_bootstrap/download/mirror_list.py",
line 63, in __init__
[fplll-5.4.1] try_lock(self.mirrorfd, LOCK_SH)  # shared (read) lock
[fplll-5.4.1]   File
"/home/dima/sage/dev/sage/build/bin/../sage_bootstrap/download/mirror_list.py",
line 34, in try_lock
[fplll-5.4.1] flock(fd, operation)
[fplll-5.4.1] KeyboardInterrupt
[fplll-5.4.1] 

[fplll-5.4.1] Error downloading fplll-5.4.1.tar.gz

and this is after "make distclean" and "rm -rf ~/.sage" - where are
these locks that are stale?!

-- 
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/CAAWYfq1LxH9jnuHHJdwrveuFCFDfFziVwQQdFrZzm5%2BUmOqEnA%40mail.gmail.com.

Re: [sage-devel] .configure stalls

2021-06-22 Thread Dima Pasechnik 
How exactly two copies of BLAS are harmful (provided that numpy always
uses its own, and it is not used by the other potential users)?
I imagine wheels mechanics is guaranteeing this - otherwise it will
break everything, not just Sage.




On Mon, Jun 21, 2021 at 6:46 PM Matthias Koeppe
 wrote:
>
> I don't think prebuilt numpy wheels can work reliably with our distribution. 
> We would be linking to two copies of BLAS - one provided by numpy, one linked 
> from our various other libraries.
> Further discussion of this - https://trac.sagemath.org/ticket/31251 
> ("Meta-ticket: Distribution as wheels").
>
> I would suggest not to try to diagnose the precise failure mode that come 
> from combining the Sage distribution with prebuilt wheels from PyPI.
>
>
>
>
> On Monday, June 21, 2021 at 9:26:26 AM UTC-7 Dima Pasechnik wrote:
>>
>> On Mon, Jun 21, 2021 at 4:45 PM Frédéric Chapoton  wrote:
>> >
>> > I did it again (see below) and it allowed the compilation of sage to start.
>> >
>> > ~/sage$ sage -pip install --upgrade --force-reinstall numpy
>> > Collecting numpy
>> > Downloading 
>> > numpy-1.20.3-cp38-cp38-manylinux_2_12_x86_64.manylinux2010_x86_64.whl 
>> > (15.4 MB)
>> > || 15.4 MB 2.9 MB/s
>> > Installing collected packages: numpy
>> > Attempting uninstall: numpy
>> > Found existing installation: numpy 1.20.3
>> > Uninstalling numpy-1.20.3:
>> > Successfully uninstalled numpy-1.20.3
>> > Successfully installed numpy-1.20.3
>>
>> I don't know how these manylinux2010 wheels are meant to work, but
>> here obviously one either had it
>> built for the dumbest possible CPU (most probably not the case here)
>> or adjust the processor capabilities during the installation,
>> something I have no idea how it could work.
>>
>> Does building sagelib error out just like before?
>>
>> >
>> > Le lundi 21 juin 2021 à 14:35:03 UTC+2, dim...@gmail.com a écrit :
>> >>
>> >> On the other hand, what did
>> >> "sage -pip install numpy" actually do? Did it fetch a binary from the net?
>> >> Did it fetch source and built?
>> >> On Monday, June 21, 2021 at 1:08:07 PM UTC+1 Dima Pasechnik wrote:
>> >>>
>> >>> On Mon, Jun 21, 2021 at 10:50 AM Frédéric Chapoton  
>> >>> wrote:
>> >>> >
>> >>> > Hello,
>> >>> >
>> >>> > So now I am back at the numpy issue when building sagelib, that I sort 
>> >>> > of "fixed" before by a forced and ugly "sage -pip install numpy".
>> >>> >
>> >>> > Is my computer just too old ? I add information about it at the end of 
>> >>> > my message.
>> >>> >
>> >>> > make --no-print-directory sagelib-SAGE_VENV-no-deps
>> >>> > cd '/home/chapoton/sage/build/pkgs/sagelib' && . 
>> >>> > '/home/chapoton/sage/src/bin/sage-src-env-config' && . 
>> >>> > '/home/chapoton/sage/src/bin/sage-env-config' && . 
>> >>> > '/home/chapoton/sage/src/bin/sage-env' && . 
>> >>> > '/home/chapoton/sage/build/bin/sage-build-env-config' && . 
>> >>> > '/home/chapoton/sage/build/bin/sage-build-env' && 
>> >>> > SAGE_SPKG_WHEELS=/home/chapoton/sage/local/var/lib/sage/wheels 
>> >>> > SAGE_INST_LOCAL=/home/chapoton/sage/local sage-logger -p 
>> >>> > '/home/chapoton/sage/build/pkgs/sagelib/spkg-install' 
>> >>> > '/home/chapoton/sage/logs/pkgs/sagelib-9.4.beta2.log'
>> >>> > [sagelib-9.4.beta2] Discovering Python/Cython source code
>> >>> > [sagelib-9.4.beta2] distributions = ['']
>> >>> > [sagelib-9.4.beta2] Discovered Python/Cython sources, time: 19.60 
>> >>> > seconds.
>> >>> > [sagelib-9.4.beta2] running build
>> >>> > [sagelib-9.4.beta2] Generating auto-generated sources
>> >>> > [sagelib-9.4.beta2] Building interpreters for fast_callable
>> >>> > [sagelib-9.4.beta2] -> First build of interpreters
>> >>> > [sagelib-9.4.beta2] running build_cython
>> >>> > [sagelib-9.4.beta2] Enabling Cython debugging support
>> >>> > [sagelib-9.4.beta2] 
>> >>> > 
>> >>> > [sagelib-9.4.beta2] Traceback (most recent call last):
>> >>> > [sagelib-9.4.beta2] File "setup.py", line 96, in 
>> >>> > [sagelib-9.4.beta2] code = setup(
>> >>> > [sagelib-9.4.beta2] File 
>> >>> > "/home/chapoton/sage/local/lib/python3.8/site-packages/setuptools/__init__.py",
>> >>> >  line 153, in setup
>> >>> > [sagelib-9.4.beta2] return distutils.core.setup(**attrs)
>> >>> > [sagelib-9.4.beta2] File 
>> >>> > "/home/chapoton/sage/local/lib/python3.8/site-packages/setuptools/_distutils/core.py",
>> >>> >  line 148, in setup
>> >>> > [sagelib-9.4.beta2] dist.run_commands()
>> >>> > [sagelib-9.4.beta2] File 
>> >>> > "/home/chapoton/sage/local/lib/python3.8/site-packages/setuptools/_distutils/dist.py",
>> >>> >  line 967, in run_commands
>> >>> > [sagelib-9.4.beta2] self.run_command(cmd)
>> >>> > [sagelib-9.4.beta2] File 
>> >>> > "/home/chapoton/sage/local/lib/python3.8/site-packages/setuptools/_distutils/dist.py",
>> >>> >  line 986, in run_command
>> >>> > [sagelib-9.4.beta2] cmd_obj.run()
>> >>> > [sagelib-9.4.beta2] File 
>> >>> > "/home/chapoton/sage/build/pkg