I would actually say that the current behavior is more useful, since it allows users to see that they are typing in a keyword name and that there will be a syntax error because of that.

I am -1 on this change, sorry :(

not total agree before this patch I tried
with ipython and other repl for python or other language none of them .xxx highlight

I am not against to stop highlighting set in a.set though :)

with ipython and other repl for python or other language none of them .xxx highlight

That's their own choices. But writing the keyword and hitting enter would anyway raise a SyntaxError so I think it's better that users know that they are writing a keyword and that they can't do that.

But I agree that the set in a.set should not be highlighted because it could be very confusing (sometimes, it's necessary to shadow builtins such as format or set)

ipython

ghci

with ipython and other repl for python or other language none of them .xxx highlight

That's their own choices. But writing the keyword and hitting enter would anyway raise a SyntaxError so I think it's better that users know that they are writing a keyword and that they can't do that.

But I agree that the set in a.set should not be highlighted because it could be very confusing (sometimes, it's necessary to shadow builtins such as format or set)

a.set and a.list the list highligt in VS Code is strange, I wonder it is the same in repl

It doesn't matter what IPython does IMO. I would prefer no highlights for built-ins when writing attributes because they can be legitimate ones, as opposed to keywords that can't be used at all. Especially with this kind of constructions:

It doesn't matter what IPython does IMO. I would prefer no highlights for built-ins when writing attributes because they can be legitimate ones, as opposed to keywords that can't be used at all. Especially with this kind of constructions:

got you point, I agreed we you do not care what IPython does, I just checked them before this patch to make sure
if I am doing something wrong

Nonetheless, while two core devs opposed changing the behavior for keywords, I'm willing to support a change of behavior when it comes to built-ins used as attributes. I think it makes sense not to highlight them.

Nonetheless, while two core devs opposed changing the behavior for keywords, I'm willing to support a change of behavior when it comes to built-ins used as attributes. I think it makes sense not to highlight them.

thank you

with ipython and other repl for python or other language none of them .xxx highlight

Just chime in to say that IPython does highlight keywords followed by whitespace:

So I agree to only keep the change in builtin names branch.

with ipython and other repl for python or other language none of them .xxx highlight

Just chime in to say that IPython does highlight keywords followed by whitespace:

So I agree to only keep the change in builtin names branch.

thank you will only leave buildin

now only not highlight the buildin words

I am not against to stop highlighting set in a.set though :)

fixed that, thank you.

I'll let REPL experts have a look in addition but thank you for the fix!

Thank you for your kindess review and @sobolevn learned a lot from this!

Thanks @yihong0618 for the PR, and @pablogsal for merging it 🌮🎉.. I'm working now to backport this PR to: 3.14.
🐍🍒⛏🤖

GH-138654 is a backport of this pull request to the 3.14 branch.