2.16 has been delayed for quite a while at this po...
# development
w
2.16 has been delayed for quite a while at this point: it’s had 7 dev releases, although in wall clock time, it’s been in development since ~Nov 1st (not terribly surprising with the holiday, but!) i think that we should consider bumping the experimental shell command stabilization to 2.17, and cutting 2.16.x this week.
@ancient-vegetable-10556: ^ thoughts?
a
If we don’t land the stabilisation this week, then fine. There’s a couple of tickets that should probably land short of renaming the targets that should be quick to implement, though
w
(or splitting what we stabilize, but that’s probably trickier)
yea, reasonable.
a
I’m definitely in favour of splitting what we stabilise though, given that’s what I’m about to propose in the discussion
🙂
that may take longer to implement than just this week, that’s all
The blockers should be #18086, #18087, and possibly #18110
f
I'd also like to fix the Go environment bug since it'll probably require removing the synthetic target stuff (which will be a win as well for pants devs in the pants repo too)
c
it’ll probably require removing the synthetic target stuff
would be interesting to preserve a note somewhere if possible about what to consider when looking at synthetic targets to avoid pitfalls (if we’d call this case such 😉 )
w
@fast-nail-55400: yea, that would be good… let us know how we can help with that.
f
I'm not likely to hit a one week deadline for it since not my day job any more. Is there going to be a alpha release?
w
yes, there will be an alpha release
s
+1 to cutting 2.16 ASAP - let me know if I can help get stuff over the line 🙂
p
I would love to sneak https://github.com/pantsbuild/pants/pull/18132 in before 2.16.0 final is released 😄 if possible...
👍 1
f
https://github.com/pantsbuild/pants/pull/18140 is an attempt to fix the non-local environment issue. It removes the need for a synthetic target, but does not actually yet fix the issue with running in a non-local (Docker in my testing) environment. Any advice on whether
EnvironmentName
or
EnvironmentField
(or some other type/rule) needs to be added in places? (The PR has a reproduction of the bug if anybody wants to take a look.)
b
(I just want 2.15.0rc3)
c
I realize I have a breaking change for the visibility backend I’d like to get in as well before we cut 2.16, to keep the break isolated in the 2.16.dev era.. for https://github.com/pantsbuild/pants/issues/17873#issuecomment-1407761256
👍 1
(not all of that comment, but the breaking bit…)
👍 1
f
okay so re the Go stuff, https://github.com/pantsbuild/pants/pull/18140 actually fixes Go running in the Docker environment apparently since the binary got built if I kept re-running
./pants package
enough times. The error I was seeing was from the Docker environment being flaky and input files not being fully visible to process being run which is flakiness in using the Docker environment, not anything particular (hopefully) to the Go backend.
also https://github.com/pantsbuild/pants/pull/18148 is necessary for Pants to work with Go v1.20 (which was just released today)
👍 1
w
@fast-nail-55400: was:
The error I was seeing was from the Docker environment being flaky and input files not being fully visible to process being run which is flakiness in using the Docker environment, not anything particular (hopefully) to the Go backend.
on macOS? if so: https://github.com/pantsbuild/pants/issues/18162#issuecomment-1421403119 … can take further discussion to the ticket or a new thread though
f
yes on macOS
also re this thread and 2.16: I've added to issues regarding bugs with the Go backend and Go v1.20: • golangci-lint tries to lint the Go SDK"coverage redesign" experiment needs to be disabled when using Go v1.20
👍 1
w
regarding the release schedule: a minor fix to make the docs match our desired reality: https://github.com/pantsbuild/pants/pull/18234