BUG: Replace on Series/DataFrame stops replacing after first NA

[asishm]

• closes BUG: DataFrame.replace with regex on StringDtype column with NA values stops replacing after first NA #56599
• Tests added and passed if fixing a bug or adding a new feature
• All code checks passed.
• Added type annotations to new arguments/methods/functions.
• Added an entry in the latest doc/source/whatsnew/vX.X.X.rst file if fixing a bug or adding a new feature.

The issue was the line a = a[mask] was triggered only for ndarray and doesn't hit when dtype='string', but the np.place logic applied as long as the result was an ndarray.

The old behavior had things like (depending on the number and location of NAs)

In [2]: s = pd.Series(['m', 'm', pd.NA, 'm', 'm', 'm'], dtype='string')

In [3]: s.replace({'m': 't'}, regex=True)
Out[3]:
0       t
1       t
2    <NA>
3       m
4       t
5       t
dtype: string

In [4]: s = pd.Series(['m', 'm', pd.NA, pd.NA, 'm', 'm', 'm'], dtype='string')

In [5]: s.replace({'m': 't'}, regex=True)
Out[5]:
0       t
1       t
2    <NA>
3    <NA>
4       m
5       m
6       t
dtype: string

In [6]: s = pd.Series(['m', 'm', pd.NA, 'm', 'm', pd.NA, 'm', 'm'], dtype='string')

In [7]: s.replace({'m': 't'}, regex=True)
Out[7]:
0       t
1       t
2    <NA>
3       m
4       t
5    <NA>
6       t
7       m
dtype: string

[asishm]

pre-commit.ci autofix

[mroeschke]

Thanks @asishm