Inconsistency in conversion from CIF and complex

[rwst]

sage: RR(CBF(1.7))
1.70000000000000

sage: CIF(1.7).is_exact()
True
sage: RR(RIF(1.7))
1.70000000000000
sage: RR(CIF(1.7))

/home/ralf/sage/src/sage/structure/parent.pyx in sage.structure.parent.Parent.__call__ (build/cythonized/sage/structure/parent.c:9530)()
    916         if mor is not None:
    917             if no_extra_args:
--> 918                 return mor._call_(x)
    919             else:
    920                 return mor._call_with_args(x, args, kwds)

/home/ralf/sage/src/sage/structure/coerce_maps.pyx in sage.structure.coerce_maps.DefaultConvertMap_unique._call_ (build/cythonized/sage/structure/coerce_maps.c:4979)()
    153                 print(type(C), C)
    154                 print(type(C._element_constructor), C._element_constructor)
--> 155             raise
    156 
    157     cpdef Element _call_with_args(self, x, args=(), kwds={}):

/home/ralf/sage/src/sage/structure/coerce_maps.pyx in sage.structure.coerce_maps.DefaultConvertMap_unique._call_ (build/cythonized/sage/structure/coerce_maps.c:4841)()
    148         cdef Parent C = self._codomain
    149         try:
--> 150             return C._element_constructor(x)
    151         except Exception:
    152             if print_warnings:

/home/ralf/sage/src/sage/rings/real_mpfr.pyx in sage.rings.real_mpfr.RealField_class._element_constructor_ (build/cythonized/sage/rings/real_mpfr.c:7305)()
    643         cdef RealNumber z
    644         z = self._new()
--> 645         z._set(x, base)
    646         return z
    647 

/home/ralf/sage/src/sage/rings/real_mpfr.pyx in sage.rings.real_mpfr.RealNumber._set (build/cythonized/sage/rings/real_mpfr.c:12976)()
   1466                 mpfr_set_inf(self.value, -1)
   1467             else:
-> 1468                 raise TypeError("unable to convert {!r} to a real number".format(s))
   1469 
   1470     cdef _set_from_GEN_REAL(self, GEN g):

TypeError: unable to convert '1.7000000000000000?' to a real number

Same with `RR(complex(1.7))`.

Component: numerical

Author: Ralf Stephan

Branch/Commit: 7b23595

Reviewer: Jeroen Demeyer

Issue created by migration from https://trac.sagemath.org/ticket/24630

[rwst]

Branch: u/rws/inconsistency_in_conversion_from_cif_and_complex

[rwst]

Commit: 1218a8a

[rwst]

New commits:

1218a8a

24630: conversions CIF-->real, Python complex-->real

[rwst]

Author: Ralf Stephan

[mezzarobba]

comment:4

Mostly looks good to me, but I don't understand why you are using parent(x) is complex instead of type(x) is complex or isinstance(x, complex).

[sagetrac-git]

Changed commit from 1218a8a to 3a41837

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

cf90ca2	Merge branch 'develop' into t/24630/inconsistency_in_conversion_from_cif_and_complex
3a41837	24630: address reviewer's comment

[mezzarobba]

comment:6

You no longer need to import parent (this should probably have been a cimport anyhow). Oh, and I missed that last time (sorry), but why are you changing RealField_class._element_constructor_() instead of RealNumber._new() — which apparently contains the logic for most other cases?

[tscrim]

comment:7

Is this issue fixed by #24371? (Also, this is to warn about a possible merge conflict.)

[rwst]

comment:8

Replying to @tscrim:

Is this issue fixed by #24371? (Also, this is to warn about a possible merge conflict.)

No, the same errors are raised.

[sagetrac-git]

Changed commit from 3a41837 to 2796d1a

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

9e5072c	Merge branch 'develop' into t/24630/inconsistency_in_conversion_from_cif_and_complex
dc476b0	Merge branch 'develop' into t/24630/inconsistency_in_conversion_from_cif_and_complex
2796d1a	24630: address reviewer's comments

[rwst]

comment:12

I just saw I didn't address your last comment, sorry.

[jdemeyer]

comment:13

Can you change the exception to

raise TypeError("f"unable to convert complex interval {self} to real number")

(just trying to be consistent and using the same wording as similar exception messages)

Also, I don't like the "Try to" in the docs. Maybe write If the imaginary part is zero, ...

Apart from that, this looks good to me.

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

7b23595

24630: address reviewer's comments

[sagetrac-git]

Changed commit from 2796d1a to 7b23595

[rwst]

comment:15

Since patchbot is okay, I'll take this as positive?

[rwst]

Reviewer: Jeroen Demeyer

[vbraun]

Changed branch from u/rws/inconsistency_in_conversion_from_cif_and_complex to 7b23595