NEP 29, bump min numpy version

Adjust %matplotlib docstring re: matplotlib_inline. (#13927) set_matplotlib_formats and set_matplotlib_close have been deprecated in IPython.display and moved out to matplotlib_inline.backend_inline; adjust the docstring of the %matplotlib magic accordingly.

Fix bug when network map is disconnected (#13925) Hi there :) I encounter a little bug with the ipython logger. When using the `%logstart` magic and writing the log to a remote path, everything works fine until the network map gets disconnected. After the network drive is disconnected you basically powerless, and you have to reopen the console. This can be easily reproduce by sharing a windows folder  Mounting it using  Start an IPython session, start a log with `%logstart Z:\log` and disconnect. Anything you will try to do from now on will not work, and you can't use the Session, and will result `OSError: [Errno 22] Invalid argument` I couldn't reproduce it in Linux environment, i guess it really depends on the `OS` here. Also weirdly this happens when the client close the connection. that's mean if the Server close the connection to the mount everything still working perfectly I added a `try except` block to allow you to stop the logger and continue using your ipython normally I Search for a way to test it but I couldn't find an easy way to reproduce it in test.

Fix formatting of usage examples for `set_env` magic (#13911) See the difference between how `env` and `set_env` look in the [docs](http://ipython.readthedocs.io/en/stable/interactive/magics.html). <img width="600" alt="image" src="https://user-images.githubusercontent.com/14007150/215095416-c8e08aca-4c61-4bc9-a84e-81209607c74f.png"> <img width="725" alt="image" src="https://user-images.githubusercontent.com/14007150/215095718-532fe30f-a4b0-446e-ba61-73c5cf1e988e.png">

Fix typos and grammar in 8.x changelog (#13918) No code changes, just updates to the changelog.

Actually install current IPython on RDT

Ignore ipynb_checkpoints

Do not use default kwarg

Improve documentation for shortcuts

Change the default for character complete to Alt+Right

Allow to customize shortcuts

Adjust %matplotlib docstring re: matplotlib_inline. set_matplotlib_formats and set_matplotlib_close have been deprecated in IPython.display and moved out to matplotlib_inline.backend_inline; adjust the docstring of the %matplotlib magic accordingly.

Fix bug when network map is disconnected

Fix typos and grammar in 8.x changelog

%time magic displays output even when code ends in semicolon #13837 (#13841) After the magic is evaluated and the result is calculated, the modification tests whether the evaluated magic was _time_ and whether semicolon is the final character. The result is killed if both things happen. My choice would be to remove the _time_ test, so a semicolon would prevent the print of the output of any magic, but this is only a suggestion I keep open. I did not write any automated test, but I can do that once (and if) the solution is accepted. [#13837](https://github.com/ipython/ipython/issues/13837) points to [#10227](https://github.com/ipython/ipython/issues/10227) (_Cell magic result in printing the last evaluated line even if followed by semicolon_). There, somebody says that ';' may be a meaningful character because we could have a C++ expression, for instance. The IPython repository says the documentation for other languages is in Jupyter. I ran Jupyter on my browser with C++ and saw that a semicolon after the last statement prevents the output to be printed (a semicolon between 2 statements in a cell seems to be necessary, though). See attached file for simple examples. Therefore, it seems that the semicolon at the end in C++ already behaves the same way that in Python and is not required by the interpreter. 

Fix disabling auto-suggestions (#13914) This should fix #13913 but I could not reproduce it in the first place. It looks like the traitlets initialization was invoking setter handler before the `__init__` had time to set `auto_suggest = None`; this is now changed so that `auto_suggest` is set to `None` on class level. It might be not as elegant because it makes it a class variable but in fact `pt_app` also uses the same pattern and since `auto_suggest` is only used on `pt_app` there is no reason not to use this pattern. Also, the type annotation was in a wrong class, I moved it.

Fix disabling auto-suggestions

Fix formatting of usage examples for `set_env` magic

back to dev

release 8.9.0