DOC: Improve slice docstrings

Update the `np.strings.slice` docstrings to follow convention
and correctly split between argument type and documentation.

Also update the `slice` release note to follow convention for
new functions.

Lastly, fix the gcc version upgrade PR reference from numpy#28101
to numpy#28102.

Author: ArvidJB

doc/release/upcoming_changes/27789.new_function.rst

@@ -1,4 +1,5 @@
-* New function `numpy.strings.slice`
-  The new function `numpy.strings.slice` was added, which implements fast
-  native slicing of string arrays. It supports the full slicing API including
-  negative slice offsets and steps.
+New function `numpy.strings.slice`
+----------------------------------
+The new function `numpy.strings.slice` was added, which implements fast
+native slicing of string arrays. It supports the full slicing API including
+negative slice offsets and steps.

doc/release/upcoming_changes/28101.change.rst renamed to doc/release/upcoming_changes/28102.change.rst

@@ -5461,20 +5461,22 @@ def add_newdoc(place, name, doc):
     Parameters
     ----------
     a : array-like, with ``StringDType``, ``bytes_``, or ``str_`` dtype
+        Input array
 
-    start : the start of the slice, an integer or an array of integers
-            which can be broadcasted to`a`'s shape
+    start : array-like, with integer dtype
+        The start of the slice, broadcasted to `a`'s shape
 
-    stop : the end point of the slice, an integer or an array of integers
-           which can be broadcasted to`a`'s shape
+    stop : array-like, with integer dtype
+        The end of the slice, broadcasted to `a`'s shape
 
-    step : the step for the slice, an integer or an array of integers
-           which can be broadcasted to`a`'s shape
+    step : array-like, with integer dtype
+        The step for the slice, broadcasted to `a`'s shape
 
     Returns
     -------
     out : ndarray
-        Output array of str or unicode, depending on input type
+        Output array of ``StringDType``, ``bytes_`` or ``str_`` dtype,
+        depending on input type
 
     Examples
     --------

numpy/_core/code_generators/ufunc_docstrings.py

@@ -1651,20 +1651,22 @@ def slice(a, start=None, stop=None, step=None, /):
     Parameters
     ----------
     a : array-like, with ``StringDType``, ``bytes_``, or ``str_`` dtype
+        Input array
 
-    start : the start of the slice, can be None, an integer or
-            an array of integers which can be broadcasted to`a`'s shape
+    start : None, an integer or an array of integers
+        The start of the slice, broadcasted to `a`'s shape
 
-    stop : the end point of the slice, can be None, and integer or
-           an array of integers which can be broadcasted to`a`'s shape
+    stop : None, an integer or an array of integers
+        The end of the slice, broadcasted to `a`'s shape
 
-    step : the step for the slice, can be None, an integer or
-           an array of integers which can be broadcasted to`a`'s shape
+    step : None, an integer or an array of integers
+        The step for the slice, broadcasted to `a`'s shape
 
     Returns
     -------
     out : ndarray
-        Output array of str or unicode, depending on input type
+        Output array of ``StringDType``, ``bytes_`` or ``str_`` dtype,
+        depending on input type
 
     Examples
     --------