#general
- c
cuddly-accountant-94274
08/05/2015, 9:33 PMOk, so what do you mean by it not being the way to go? - c
cuddly-accountant-94274
08/05/2015, 9:33 PMUnless you just mean "wait for python 2 support" - c
cool-thailand-74745
08/05/2015, 9:33 PMyes, that - c
cuddly-accountant-94274
08/05/2015, 9:33 PMOk. - c
cool-thailand-74745
08/05/2015, 9:34 PMbasically i mean -- don't do the annotation stuff - c
cool-thailand-74745
08/05/2015, 9:34 PMjust go the .pyi route - c
cool-thailand-74745
08/05/2015, 9:34 PMwhich could be a pretty straightforward extension to pants - c
cuddly-accountant-94274
08/05/2015, 9:34 PMAre you proposing use of .pyi for new code, not just for adding type hints to existing code? - c
cuddly-accountant-94274
08/05/2015, 9:35 PMIt's purpose seems to be to allow retro-active hinting without original code modification. - c
cool-thailand-74745
08/05/2015, 9:35 PMexisting and new code - c
cool-thailand-74745
08/05/2015, 9:35 PMi'm just saying, don't tie us to py3 simple smile - r
rough-minister-58256
08/05/2015, 9:36 PM@cuddly-accountant-94274 is this for your RB rewrite? - c
cuddly-accountant-94274
08/05/2015, 9:36 PMThat seems to criple it quite a lot, having to maintain a separate file with type hints. - c
cuddly-accountant-94274
08/05/2015, 9:36 PMWell that would be one example. I want to be able to write maintainable non-trivial code in a language suitable for tooling. - c
cuddly-accountant-94274
08/05/2015, 9:36 PMAnd make it easier for different people to collaborate, refactor, etc. - c
cool-thailand-74745
08/05/2015, 9:37 PMit would be great for .pyi annotations to be part of docstrings😕 1 - c
cool-thailand-74745
08/05/2015, 9:37 PMbut that would be too logical for pep484 - c
cuddly-accountant-94274
08/05/2015, 9:38 PMI'll have to read PEP-484 more carefully - c
cuddly-accountant-94274
08/05/2015, 9:38 PMBut my knee jerk reaction is: Those type hints should be obvious to the reader of the code, not just the type checking system. - r
rough-minister-58256
08/05/2015, 9:38 PMwhats the argument against going all-in on py3+type hinting for new projects? - c
cuddly-accountant-94274
08/05/2015, 9:39 PMI actually wonder as well - is it just that we are fraid of python 2 lingering, or do we actively not want to adopt python 3 for some reason? - c
cuddly-accountant-94274
08/05/2015, 9:39 PMOther than "it's hard to switch", Python 3 to be is one of those no-brainers in terms of trying to get on board. But migration is obviously painful. - r
rough-minister-58256
08/05/2015, 9:39 PMwe could isolate py3 CI under a separate ‘all3’ target or something to solve the current test isolation pains - w
witty-crayon-22786
08/05/2015, 9:42 PMat a base level, incompatibilities make it harder to share code - w
witty-crayon-22786
08/05/2015, 9:42 PMpart of the reason for building from source is to remove that barrier - c
cuddly-accountant-94274
08/05/2015, 9:42 PMBut the argument here was: Don't tie us to Python 3. - c
cuddly-accountant-94274
08/05/2015, 9:42 PMNot "don't mix versions". - r
rough-minister-58256
08/05/2015, 9:42 PMmakes sense - c
cuddly-accountant-94274
08/05/2015, 9:42 PMAnd I'm wondering why. - r
rough-minister-58256
08/05/2015, 9:43 PM@cuddly-accountant-94274, tho you will be charting new territory in terms of infrastructure support - afaict, we don’t even have py3 packaged for prod etc