That worked like a charm!
* frame #0: 0xffff83c0f9024900
frame #1: 0x0000000153ff8b7c _dop.cpython-39-darwin.so`PyFortranObject_New
+ 40
frame #2: 0x0000000153ff66fc _dop.cpython-39-darwin.so`PyInit__dop +
1020
And it turns out that this problem was independently reported by William
Stein. I used an external gfortran compiler installed in /usr/local to
build this sage. So I need to also embed the fortran runtime library into
my Sage.framework and patch the load paths of all executable files that
load it. This would be a lot easier if someone could get the gfortran spkg
to build on an M1 CPU.
Anyway, thanks! I think I am on track to make this work now.
If anyone knows what dop stands for, I am all ears.
- Marc
On Mon, Jan 17, 2022 at 5:35 PM Matthias Köppe <[email protected]>
wrote:
> Use "./sage -sh" and run python / the debugger from there?
>
> On Monday, January 17, 2022 at 2:35:38 PM UTC-8 [email protected] wrote:
>
>> I get the same exception on both Intel and Arm when I try to import
>> sage.all from the sage python. There is no exception if I do the same
>> import in sage -python. But that is no help in terms of running lldb.
>> However, perhaps if I knew how to set up the correct environment first I
>> would be able to start the sage python in lldb and then actually be able to
>> do the import.
>>
>> - Marc
>>
>>
>> On Monday, January 17, 2022 at 3:18:05 PM UTC-6 [email protected]
>> wrote:
>>
>>> Try "import sage.all" first.
>>>
>>> On Monday, January 17, 2022 at 1:13:38 PM UTC-8 [email protected]
>>> wrote:
>>>
>>>> While many things work in the binary macOS app for Arm CPU, I
>>>> discovered by accident that the following produces a segfault:
>>>>
>>>> sage: M = Manifold()
>>>>
>>>> In fact, even trying to import manifold from sage.manifolds segfaults.
>>>> This does not happen on the Intel version of the app built from the same
>>>> branch of the repository.
>>>>
>>>> I have not been able to run a debugger on sage at all. Of course
>>>> cysignals cannot produce a backtrace on macOS. With lldb -p I get an error
>>>> saying that lldb is not allowed to attach to the process and referring me
>>>> to non-existent messages in system.log for explanations. Running the
>>>> debugserver directly produces:
>>>> error: failed to attach process 6619: unable to start the exception
>>>> thread
>>>>
>>>> This occurs even if I run sage directly in the repo after the build
>>>> finishes - it does not have anything to do with the app. (And the debugger
>>>> similarly refuses to attach to the process.) I tried to run the debugger
>>>> as root and I also tried adding debugger entitlements and resigning the
>>>> app. Neither make it possible for me to use lldb on sage itself.
>>>>
>>>> I am able to start the sage python3 in lldb. If I knew how to import
>>>> the Manifold class from python I might be able to get somewhere. But if I
>>>> run
>>>> >>> from sage.manifolds import manifold
>>>> I get a python exception whose traceback ends with:
>>>> File "sage/structure/category_object.pyx", line 60, in init
>>>> sage.structure.category_object
>>>> (build/cythonized/sage/structure/category_object.c:9914)
>>>> ImportError: cannot import name Category
>>>>
>>>> - Marc
>>>>
>>>> On Saturday, January 15, 2022 at 4:30:08 PM UTC-6 Marc Culler wrote:
>>>>
>>>>> Hi Matthias,
>>>>>
>>>>> I have posted binary releases for 9.5.rc1 on macOS with Intel or Arm
>>>>> CPUs at
>>>>> https://github.com/3-manifolds/Sage_macOS/releases/tag/v1.3-rc1.
>>>>> There are two separate disk images. You may recall that I do not use
>>>>> Homebrew or Conda, but I did have to use an experimental binary release of
>>>>> gfortran for M1 since I was not able to build the gfortran spkg.
>>>>>
>>>>> - Marc
>>>>>
>>>>> On Saturday, January 15, 2022 at 3:01:14 PM UTC-6 [email protected]
>>>>> wrote:
>>>>>
>>>> Based on the runs on GH Actions, I have updated the summary of platform
>>>>>> support in https://wiki.sagemath.org/ReleaseTours/sage-9.5#Sources
>>>>>>
>>>>>> TL;DR: It's all good now on all platforms tested on GH Actions,
>>>>>> including Cygwin.
>>>>>> But I would urge that 9.5 still merge the positively reviews tickets
>>>>>> marked "critical"/"blocker", such as
>>>>>> https://trac.sagemath.org/ticket/29285 (which fixes the build in the
>>>>>> presence of pyenv).
>>>>>> Also *note we have absolutely no automatic testing for the Apple
>>>>>> Silicon (M1) platform.* So it would be good if users/developers who
>>>>>> have access to this platform test the release candidate.
>>>>>>
>>>>>>
>>>>>>
>>>>>> On Friday, January 14, 2022 at 8:14:13 AM UTC-8 Matthias Köppe wrote:
>>>>>>
>>>>>>> On Thursday, January 13, 2022 at 3:22:03 PM UTC-8 Volker Braun wrote:
>>>>>>>
>>>>>>>> As always, you can get the latest beta version from the "develop"
>>>>>>>> git branch. Alternatively, the self-contained source tarball is at
>>>>>>>> http://www.sagemath.org/download-latest.html
>>>>>>>>
>>>>>>>> 3da1b22c25 (HEAD -> develop, tag: 9.5.rc1, trac/develop) Updated
>>>>>>>> SageMath version to 9.5.rc1
>>>>>>>>
>>>>>>>
>>>>>>> Thanks for the new rc. Tests are running at
>>>>>>> https://github.com/sagemath/sage/actions/runs/1690315877 (Linux,
>>>>>>> macOS) and https://github.com/sagemath/sage/actions/runs/1690315874
>>>>>>> (Cygwin).
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>> --
> You received this message because you are subscribed to a topic in the
> Google Groups "sage-release" group.
> To unsubscribe from this topic, visit
> https://groups.google.com/d/topic/sage-release/8ctwsUGl6LQ/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to
> [email protected].
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/sage-release/8124ca5a-7893-4f48-b3cc-af37c51d144cn%40googlegroups.com
> <https://groups.google.com/d/msgid/sage-release/8124ca5a-7893-4f48-b3cc-af37c51d144cn%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>
--
You received this message because you are subscribed to the Google Groups
"sage-release" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/sage-release/CALcZXRGQNX3H-b07u4Ph2FLAKeVV4JgdtZwP649_kH0fc3MKXw%40mail.gmail.com.
