Sorry to think out loud, but if the lexer marked `foo` as a generic `Word` token, that could be a keyword or a name, then the parser could look at the value of each `Word` token, and if the context is `import foo`, `class foo...`, `def foo...` or `foo = ...`, then `foo` is a name there and thereafter (and `x foo y` is a SyntaxError) else... you get the idea.
-- Carl Smith carl.in...@gmail.com On 14 May 2018 at 22:06, Carl Smith <carl.in...@gmail.com> wrote: > Just to be clear, if `foo` was introduced as a new infix operator, > projects that used `foo` > as a name would not be able to also use `foo` as an infix operator in the > file that defines > `foo` as a name, but could use the operator throughout the rest of their > project. > > -- Carl Smith > carl.in...@gmail.com > > On 14 May 2018 at 21:52, Carl Smith <carl.in...@gmail.com> wrote: > >> I can only think of three ways to reference a name defined in a different >> file: In an import >> statement, as properties of objects and as keyword arguments. >> >> Import statements are implicit assignments, so if Python allowed the >> following grammar, >> you could still import the odd thing that had a reserved name, without >> bringing that name >> into your local namespace. >> >> from <keyword> import <keyword> as <name> >> >> Property names always follow a dot, where only a name is valid, so Python >> could allow >> this too: >> >> <expression>.<keyword> >> >> Keyword arguments are also generally unambiguous, as they have to appear >> within the >> parens of an invocation, before the equals sign: >> >> <expression>(<keyword>=<expression>) >> >> If Python allowed those three examples (but still prevented users from >> *defining* names >> that are keywords) new keywords could be introduced without breaking old >> code , and the >> language would only require relatively minor tweaking. >> >> -- Carl Smith >> carl.in...@gmail.com >> >> On 14 May 2018 at 19:11, Terry Reedy <tjre...@udel.edu> wrote: >> >>> On 5/14/2018 10:02 AM, Clint Hepner wrote: >>> >>>> >>>> On 2018 May 14 , at 6:47 a, Daniel Moisset <dmois...@machinalis.com> >>>>> wrote: >>>>> >>>>> Following up some of the discussions about the problems of adding >>>>> keywords and Guido's proposal of making tokenization context-dependent, I >>>>> wanted to propose an alternate way to go around the problem. >>>>> >>>> >>>> My main objection to what follows is that it doesn't seem to offer any >>>> benefit over the current practice of appending an underscore (_) to a >>>> keyword to make it a valid identifier. >>>> >>> >>> Tkinter uses this convention for a few option names that clash. >>> >>> -- >>> Terry Jan Reedy >>> >>> >>> _______________________________________________ >>> Python-ideas mailing list >>> Python-ideas@python.org >>> https://mail.python.org/mailman/listinfo/python-ideas >>> Code of Conduct: http://python.org/psf/codeofconduct/ >>> >> >> >
_______________________________________________ Python-ideas mailing list Python-ideas@python.org https://mail.python.org/mailman/listinfo/python-ideas Code of Conduct: http://python.org/psf/codeofconduct/
