Re: [WIP] New package: sagemath-9.5.beta8

[tornaria <tornaria@users.noreply.github.com>]

[-- Attachment #1: Type: text/plain, Size: 2619 bytes --]

New comment by tornaria on void-packages repository

https://github.com/void-linux/void-packages/pull/34030#issuecomment-1000466098

Comment:
Works for me on glibc and also on a xbps-src chroot with musl libc, in fact all tests pass on x86_64 and x86_64-musl.
```
$ ./xbps-src -m masterdir.musl -Ct chroot
[...]
[.../masterdir.musl /]$ /usr/lib/sage-9.5.beta8/sage 
┌────────────────────────────────────────────────────────────────────┐
│ SageMath version 9.5.beta8, Release Date: 2021-12-12               │
│ Using Python 3.10.1. Type "help()" for help.                       │
└────────────────────────────────────────────────────────────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃ Warning: this is a prerelease version, and it may be unstable.     ┃
┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛
sage: V = var('r1, r2, u, v, m')
sage: r1 = exp(m*(u^2-v^2))*u
sage: r2 = exp(m*(v^2-u^2))*v
sage: Jac = jacobian((r1,r2),(u,v)).inverse().simplify_full()
sage: Jac
[(2*m*v^2 + 1)*e^(m*v^2)/(2*m*v^2*e^(m*u^2) + (2*m*u^2 + 1)*e^(m*u^2))                     2*m*u*v*e^(m*u^2 - m*v^2)/(2*m*u^2 + 2*m*v^2 + 1)]
[      2*m*u*v*e^(m*v^2)/(2*m*v^2*e^(m*u^2) + (2*m*u^2 + 1)*e^(m*u^2))               (2*m*u^2 + 1)*e^(m*u^2 - m*v^2)/(2*m*u^2 + 2*m*v^2 + 1)]
```

Did you manage to compile with debug symbols to see if we can get a better stack trace?

Does this happen always exactly the same, at the same point? A segmentation fault is very weird at that point...

As for the jupyterlab stuff... for now it's bound to be inconvenient. Sage installs python modules in its own venv which is isolated from system site-packages, but now that system python 3.10 is working we'll eventually try https://trac.sagemath.org/ticket/29665 which will allow us to configure the sage venv to default to system site-packages so we can use system python modules.

We are really close to having a working modular sagemath. Then we can start to think about the venv and integration.