Python 3.13 Debuts With New Interactive Interpreter & Experimental JIT

[phoronix]

Phoronix: Python 3.13 Debuts With New Interactive Interpreter & Experimental JIT

Following a last minute delay due to a performance regression, Python 3.13 stable is out today as the annual major feature release to this widely-used scripting language implementation...

https://www.phoronix.com/news/Python-3.13-Released

[andreano]

New REPL: I heard it supports multiline editing (yay!), but what's new compared to ipython? Did they just rename ipython? That's at least the first thing I would do.

[fitzie]

New REPL: I heard it supports multiline editing (yay!), but what's new compared to ipython? Did they just rename ipython? That's at least the first thing I would do.

never really got used to ipython, I'm more of a pdb/pdbpp/pdbp user. I thought most ipython people moved to jupyter these days.

[habilain]

New REPL: I heard it supports multiline editing (yay!), but what's new compared to ipython? Did they just rename ipython? That's at least the first thing I would do.

It's actually based on the PyPy REPL, so not ipython based.

[uid313]

New REPL: I heard it supports multiline editing (yay!), but what's new compared to ipython? Did they just rename ipython? That's at least the first thing I would do.

I just tried it and the new REPL is better than the old REPL because it has colors and points out where on the line it ran into a error. It is still not as good as IPython though, because IPython applies syntax highlighting as you type while the Python REPL doesn't. Both IPython and Python REPL provides auto-completion but IPython has much better auto-completion because it shows in gray the characters that it can auto-complete with, and it discovers more things that it can autocomplete by looking in other modules, and it can even autocomplete modules to import when you type import. IPython can also run commands such as ls and cd, etc.

[rmfx]

I would have preferred that PyPy is adopted as the future official implementation instead, and compliance and features to be worked on before it's fully replacing CPython.

[PuckPoltergeist]

I would have preferred that PyPy is adopted as the future official implementation instead, and compliance and features to be worked on before it's fully replacing CPython.

PyPy is at python3.10, so they had to catch up three versions to replace cpython. Additionally PyPy is not fully compatible to cpython, so this has to be addressed too. And last but not least, the GIL affects PyPy too. All in all it seems to be the easier way to do the work in cpython.

[habilain]

I would have preferred that PyPy is adopted as the future official implementation instead, and compliance and features to be worked on before it's fully replacing CPython.

Not likely to happen; a major use case of Python is as a glue language (i.e. to stick different components together), and PyPy is not completely compatible in this regard. Further, PyPy has a large maintenance burden, with compilation being a... rather long process, and is written in a dialect of Python 2.

[Gamer1227]

Not likely to happen; a major use case of Python is as a glue language (i.e. to stick different components together), and PyPy is not completely compatible in this regard. Further, PyPy has a large maintenance burden, with compilation being a... rather long process, and is written in a dialect of Python 2.

Yep, and if you more performace, just put static types on your codd with the Typing module, and run mypc to compiled typed code to C code.

It will give a 2-5x performace boost, and is easier than using a non-compatible implementation. (The speed boost is only for non-numeric code, for numeric code you should just use Numpy)