Real/Complex number str() method: do not truncate by default and allow specifying number of digits

[jdemeyer]

1. Use truncate=False by default in the str() method but use truncate=True in __repr__. For explicit string conversions using the str() method, I assume that the user wants full control and doesn't want random truncation to interfere. This can also be seen in the Sage library: there are a lot more places which call x.str(truncate=False) than x.str(truncate=True). For normal printing, or conversion to str, truncation is still the default.

2. Implement a x.str(digits=20) argument to explicitly choose the number of digits for printing a real number.

[leif]

Do I have to look into the doctesting framework or will this ticket just fix #21121 as a by-product?

[jdemeyer]

It will not fix #21121 as by-product, but #21121 will be trivial to fix using #21124.

[jdemeyer]

New commits:

​e8816a8	Avoid a compiler warning
​a2933d9	Improve RealNumber.str()

[leif]

--- a/src/sage/rings/complex_mpc.pyx
+++ b/src/sage/rings/complex_mpc.pyx
@@ -1026,32 +1026,32 @@ cdef class MPComplexNumber(sage.structure.element.FieldElement):
- ``base`` -- base for output
- - ``truncate`` -- if ``True``, round off the last digits in printing
- to lessen confusing base-2 roundoff issues.
+ - ``**kwds`` -- other arguments to pass the the ``str()``
+ method of the real numbers in the real and imaginary parts.
EXAMPLES::

s/the the/to the/

--- a/src/sage/rings/complex_number.pyx
+++ b/src/sage/rings/complex_number.pyx
@@ -465,17 +465,19 @@ cdef class ComplexNumber(sage.structure.element.FieldElement):
- ``base`` -- (Default: 10) The base to use for printing
- - ``truncate`` -- (Default: ``True``) Whether to print fewer
- digits than are available, to mask errors in the last bits.
-
- ``istr`` -- (Default: ``I``) String representation of the complex unit
+ - ``**kwds`` -- other arguments to pass the the ``str()``
+ method of the real numbers in the real and imaginary parts.
+
EXAMPLES::

s/the the/to the/

I.e., sed -i 's/ the the / to the /' src/sage/rings/complex_{mpc,number}.pyx

There's some inconsistency in giving the defaults of arguments:

(Default: 10) vs. (default:10) or even just mentioning it in the text (e.g. for no_sci, where the default is None), or in some cases also for truncate, where the default now is False, but the description says (just) "if True, ...".

(But that's a minor thing, as you'd presumably have to touch more regions to normalize all instances.)

P.S.: I looked at your previous commit, and just briefly scanned the new one.

[git]

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

​1de12e3

Minor doc fixes

[git]

Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:

​6955cfe

Improve RealNumber.str()

[jdemeyer]

Rebased to 7.6.rc2

[git]

Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:

​1993e82

Improve RealNumber.str()

[tscrim]

LGTM.