On Thursday, 1 June 2023 at 07:29:27 UTC-7 Luis Finotti wrote:
I am not sure I will miss them (and other packages that depended on it -- fflas-ffpack libgivaro-dev libgivaro9 liblinbox-dev python3-brial python3-sage), but isn't there a way to force Jupyter to use Sage's one? I don't think jupyter cares about givaro. However, it sounds like your jupyter sets up some things about the library search path (perhaps an LD_LIBRARY_PATH)? that gets inherited by the sage process that gets subsequently run to start the kernel. The sage script normally seems to set search paths in such a way that it picks the right libgivaro, but when run in the environment created by jupyter apparently it doesn't any more. That sounds like a bit of a bug to me. You've found a work-around by making sure the only givaro on your system is the one sage built. But really, the way sage sets up its environment shouldn't be derailed by what another jupyter install does prior to it. The first candidate for causing this is a value for LD_LIBRARY_PATH, so if you could look check "LD_LIBRARY_PATH" in os.environ in both your command-line sage and in sage through jupyterlab then you'd know if there's a difference. Generally comparing the contents of os.environ between the two may give you some insight in what might be causing the difference in behaviour. The whole point of python's "venv" is that you CAN run different programs with conflicting library requirements easily. So you could help if you can pinpoint what is going wrong here -- perhaps it's a misconfiguration on your system but it could also be a bug/oversight in how python's venvs are managed. -- You received this message because you are subscribed to the Google Groups "sage-support" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-support+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/sage-support/032ca71d-5f67-44db-89dc-ed0423201135n%40googlegroups.com.
