Why not a const keyword?

[Lucas_Malor]

Since in these days there’s a major effort to make CPython faster, I think this is of some interest.

I’m sure the introduction of a const keyword, also with other names, has been raised many times. And I’m sure it was every time discarded.

My question is: why not a const keyword? Too much difficult to implement in actual ecosystem? Too low speed gain? Anything else?

[bschubert (Brian Schubert)]

Some previous related discussions:

Compiler-Level const-Optimization

Python-Interpreter does not check for unused (or write only) variables, and does introduce unnecessary variables (local values). At least for numbers (integer and float) it should be possible to improve byte-code-level (regarding performance and efficency). I want to propose a new pseudo-type “const” (acting like C-defines). Also “const” auto-detection on complier/interpreter-level to create better byte-code improves here. Skipping Write-Once-Read-Immediately temp-values. # check Interpreter …

Adding constant variables

In many other languages, you can create a constant variable, one that does not change, typically using const or constant. In Python, it is common for the name of constants to use an uppercase name (PI = 3.14 for example), but these are still changeable. I wouldn’t know, since I haven’t gone too deep with the C behind Python, but it shouldn’t be hard to implement, or you could add a const keyword to explicitly state the fact that a variable is constant. EDIT: I came up with a base idea for how …

Mailman 3 Introduce constants in Python (constant name binding) -...

Mailman 3 [RFC] "Strict execution mode" (TL;DR version) - Python-ideas -...

Mailman 3 Re: [Python-ideas] Should Python have user-defined constants? -...

Should constants be introduced to Python?

Mailman 3 'const' and 'require' statements - Python-ideas - python.org