3 ms·
Yes, I could do that. But as it is, I'm no worse off than anyone else with Python2 codebases. I used to use the future imports and the things you're suggesting
by BuckRogers 11y ago
Yes, I could do that. But as it is, I'm no worse off than anyone else with Python2 codebases.
I used to use the future imports and the things you're suggesting but since my job is Python2.x I stopped. I prefer to just keep my head there all the time. If I were to migrate to CPython3 instead of PyPy4 as I've done, it would be a wholesale move. Employer moves, I move. The underlying incentives aren't there though at the moment for either of us.
I'm certainly not going to push my employer into something that isn't in their best-interests (or mine, work for nothing). If anything, they need to move to PyPy for the same reasons I did.
Py3 needs a bigger ecosystem than 2.x, needs an answer to PyPy4's CPU performance and removal of the GIL, and convince employers it's better than CPython2 and/or PyPy4.
Those are 4 potentially insurmountable tasks.
The best thing to do is start removing features rather than adding them, such as this 4th string formatting method in 3.6. It (Python3) is just very unappealing.
- rufugee 11y agoIs there something fundamentally different from Py3 that makes a well-performing functional PyPy version impossible or extremely difficult?
- BuckRogers 11y agoLate reply here, but no. The reason there's no non-beta PyPy3 on a later version than 3.2 is that there's not enough demand for it (or no one willing to donate large enough sums of money for it to make it the primary focus). Industry players who want to remove the Python CPU bottleneck and willing to stick with Python rather than moving on to Go or JS on Node, are donating to the 2.7 based PyPy4.