back to dev

Fix release

release 9.4.0

Add changelog for IPython 9.4 (#14935) Add changelog for IPython 9.4

Align formatting of section references with one used in 8.x series

Mention `%script` magic improvements

Add changelog for IPython 9.4

Do not always import * with autoreload3 (#14872) Fixes #14839 Previously, `%autoreload 3` always had the effect of `from module import *` on reloads. This PR fixes this functionality to the expected utility: if users specifically import some symbols from a module (for example `from X import Y` or `from X import Y as Z`), only those symbols are reloaded. Unit tests are added to demonstrate the new behavior as well.

Backport Pdb closure evaluation fix (list comprehension, generators) from Python 3.13 (#14933) - [x] add a failing test case - [x] backport https://github.com/python/cpython/pull/111094 into IPython codebase - fixes https://github.com/gotcha/ipdb/issues/256 - solves the pdb-related comments on https://github.com/ipython/ipython/issues/62  ### Context IPython already includes bits of Pdb code from Python itself, as documented in: https://github.com/ipython/ipython/blob/1e9a2e8b5a584ffe1973318fdabf2e39395ec367/IPython/core/debugger.py#L92-L110 I am taking a more cautious approach and separating the backport to a new file to make it easy to delete in the (near) future (1 year from now).

Add mypy ignore comments

Lint with darker (black)

Include a backport of Pdb closure fixes from Python 3.13

Add a test for closure evaluation

Fix streaming multi-byte Unicode characters in `%script` magic family (#14931) - [x] Adds a test case - [x] Fixes #14930

Fix the test on Windows

Fix the issue

Fix test (do not print out a new line)

Add a test case

Fix access, Darker

Support data frames, series and objects with `__len__` in the `%whos` magic (#14926) This modification shows the shape of dataframe and series objects instead of printing the dataframe/series objects. ```python ## With this slight modification [1] a,b,c = 1,2,3 [2] df = pd.DataFrame({"a":range(10),"b":range(10,20)}) [3] s = df["a"] [4] %whos Variable Type Data/Info --------------------------------- a int 1 b int 2 c int 3 df DataFrame Shape: (10, 2) s Series Shape: (10,) ``` Compare the above to the way it currently prints the output: ```python ## current way [1] a,b,c = 1,2,3 [2] df = pd.DataFrame({"a":range(10),"b":range(10,20)}) [3] s = df["a"] [4] %whos Variable Type Data/Info --------------------------------- a int 1 b int 2 c int 3 df DataFrame a b\n0 0 10\n1 <...>\n\n[10 rows x 2 columns] s Series 0 0\n1 1\n2 2\n <...> Length: 10, dtype: int64 ```