-
Notifications
You must be signed in to change notification settings - Fork 70
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
drop Go 1.16, add Go 1.18 #158
Conversation
While here, bump gotip and the action versions.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM with one thought. WDYT?
@@ -1,5 +1,5 @@ | |||
module github.com/rogpeppe/go-internal | |||
|
|||
go 1.16 | |||
go 1.17 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Could we leave support for go1.16 until we actually require go1.17 features?
I suspect there are a bunch of people still using 1.16 out there.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Is that a policy we want to enforce going forward? We didn't support Go 1.15 when 1.17 was the latest version, so what you say would be a change - supporting 3 major versions rather than 2. I personally think it's unnecessary at this point, given that 1.18 has been out for three months and 1.16.x has been unsupported upstream for a while too :)
Worth noting that bumping this version here doesn't preclude building on 1.16; in fact it should continue to work until we actually do start using new features/APIs. But more than anything, it's a signal that we only test and explicitly support 1.16 or later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I reckon that it's nicer to keep support for a given version until we need to drop it tbh. Is there any significant cost to us doing so?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
No immediate cost, but CI does get slower the more versions we support, and we do have shenanigans like dealing with the signature of testing.MainStart
for example. We have 6 CI jobs and we'd likely jump to 9.
If we want to change the supported version policy that's fine with me, I just wasn't aware that was your intention. Either way, I'll leave it to you to wrap up this PR however you see fit, because I don't have strong opinions either way :)
(see commit message)