We are heavily dependent on V8, which in turn is heavily dependent on Chromium tooling, which does not support Python3 and has only recently begun conversion -
https://bugs.chromium.org/p/chromium/issues/detail?id=942720
https://bugs.chromium.org/p/v8/issues/detail?id=8594
https://github.com/nodejs/node/issues/24512
I'm happy to do this, I think it'll be straightforward, but it's probably worth waiting for v8 support to be completed.
it's probably worth waiting
Not! Python 3 was released 11 years ago and there are 113 days until Python 2 end of life. Please do work in _parallel_, not serial to be ready before yearend. It would be great if someone could get v8/v8#33 and v8/v8#34 pushed and landed upstream.
@cclauss I bet I can do the deno part of the port in under an hour, certainly less than a day. I've ported large modules to python 3... deno is very small. I suppose it would help to enumerate precisely which python v8 scripts need to be ported in order for deno to work, I can do that.
Not a question of your skills. I have been impressed. It is just a question of putting lots of concentrated pressure on the few bottlenecks that are not ready so we create a _compelling reason to act_. We want to create the impression that the train is leaving the station without the laggards onboard.
Hi!
I send this PR from @AgentJ08 to https://chromium-review.googlesource.com/c/v8/v8/+/1864942
https://travis-ci.com/bmsdave/v8/builds/132539826
Does this mean v8 can be built with python 3?
@bmsdave Can V8 be built with Python 3?
https://travis-ci.com/bmsdave/v8/builds Seems to be passing the lint steps. Where does the V8 build on Py3 break down?
The script it's running is true, the Travis build is not running any build/tests other than flake8...
Exactly. So where can we see the __build__ failing on Python 3?
Sorry I was answering (no) to @kt3k :
https://travis-ci.com/bmsdave/v8/builds/132539826
Does this mean v8 can be built with python 3?
It could be that the build works with just the syntax changes (but I would be surprised).
I think we need to compile and run with gm.py, see https://v8.dev/docs/test.
Closing in favour of #8223 (plus we have #7779).
Most helpful comment
@cclauss I bet I can do the deno part of the port in under an hour, certainly less than a day. I've ported large modules to python 3... deno is very small. I suppose it would help to enumerate precisely which python v8 scripts need to be ported in order for deno to work, I can do that.