py3: unhandled exception in pynac

[embray]

There are several bugs like:

sage: arccoth(float(1.1))
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
AttributeError: type object 'float' has no attribute 'precision'

The above exception was the direct cause of the following exception:

SystemError                               Traceback (most recent call last)
<ipython-input-1-96f50dbce315> in <module>()
----> 1 arccoth(float(RealNumber('1.1')))

/home/embray/src/sagemath/sage-python3/local/lib/python3.6/site-packages/sage/symbolic/function.pyx in sage.symbolic.function.BuiltinFunction.__call__ (build/cythonized/sage/symbolic/function.cpp:11828)()
    993             res = self._evalf_try_(*args)
    994             if res is None:
--> 995                 res = super(BuiltinFunction, self).__call__(
    996                         *args, coerce=coerce, hold=hold)
    997

/home/embray/src/sagemath/sage-python3/local/lib/python3.6/site-packages/sage/symbolic/function.pyx in sage.symbolic.function.Function.__call__ (build/cythonized/sage/symbolic/function.cpp:6847)()
    487             res = g_function_evalv(self._serial, vec, hold)
    488         elif self._nargs == 1:
--> 489             res = g_function_eval1(self._serial,
    490                     (<Expression>args[0])._gobj, hold)
    491         elif self._nargs == 2:

SystemError: <built-in function __import__> returned a result with an error set

This is due to an unhandled exception elsewhere in numeric.cpp. This is a problem on Python 2 as well, but by luck the exception is cleared later before it becomes a problem (whereas Python 3 is more aggressive in checking for unhandled exceptions when going into a function call).

Upstream PR: pynac/pynac#307

Upstream: Fixed upstream, in a later stable release.

CC: @rwst

Component: python3

Keywords: pynac

Issue created by migration from https://trac.sagemath.org/ticket/24752

[jdemeyer]

comment:1

Typically, pynac releases quickly after a bug-fix so we should wait for that.

[jdemeyer]

Upstream: Reported upstream. No feedback yet.

[embray]

comment:2

I know, but it's still worth noting that this issue can be changed after the next pynac upgrade.

[embray]

Changed upstream from Reported upstream. No feedback yet. to Fixed upstream, but not in a stable release.

[jdemeyer]

Dependencies: #24745

[embray]

comment:4

I believe the pynac update in #24745 will resolve several other Python 3 issues I've already fixed, but I guess not this one?

[rwst]

comment:5

Yes, that one is completely different.

Usually we use bug report tickets that are fixed in Pynac for adding doctests but, since it came from a doctest, it's practically a duplicate of a yet-to-be-opened Pynac upgrade ticket. Alternatively, you could add your PR as patch here. The Pynac upgrade will then just remove it.

[fchapoton]

Changed upstream from Fixed upstream, but not in a stable release. to Fixed upstream, in a later stable release.

[fchapoton]

Changed dependencies from #24745 to none