[Discussion] lintr 3.0.0 release?

[AshesITR]

We've added a lot of new and exciting features to the current development version of lintr, to the point that some downstream users (including me) are using the github version because they need e.g. directory exclusions.

So I wonder, is it time for prepairing a new release of lintr?
I think the number of changes warrants a major version bump, so I suggest 3.0.0 to be the new version.

[MichaelChirico]

definitely agree we should try and triage open PRs/issues and push for a release soon. we definitely need to bump minor version, I don't have a good sense of whether major version should be bumped.

…

On Fri, Jan 29, 2021, 12:00 PM AshesITR ***@***.***> wrote: We've added a lot of new and exciting features to the current development version of lintr, to the point that some downstream users (including me) are using the github version because they need e.g. directory exclusions. So I wonder, is it time for prepairing a new release of lintr? I think the number of changes warrants a major version bump, so I suggest 3.0.0 to be the new version. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#718>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB2BA5L5JQIBNPS5TA75KR3S4MHW5ANCNFSM4WZNLHSQ> .

[AshesITR]

I looked at the current NEWS.md and found two changes, that might impact "off-label" usage of lintr:

• exclusions are now relative to the .lintr file (they were relative to the caller wd before that):
  this could break code where a user conventionally calls e.g. lint_package() from outside the package and adapted the .lintr config to the previous (IMO quirky) behaviour
• lints now have the linter name instead of the implementation name assigned:
  this could break contrived code where the names in linters are specified and later the Lint names used with the assumption that they are equal to the implementing functions, e.g. linters = list(a = line_length_linter()) produced lints from the line_length_linter while it now produces lints from the a linter. This seems unlikely to have relevance in practice

[MichaelChirico]

Went ahead and created a milestone (3.0.0) to tag issues/PRs we'd like to close out for release.

[MichaelChirico]

LMK when we're about ready to freeze new features, I'll start a comparison of HEAD vs 2.0.1 on all CRAN

[AshesITR]

I think the most important open issue is #720 atm. If we agree to settle on the current implementation, I would try finishing all open milestone:3.0.0 PRs and start with release preparations.

[MichaelChirico]

Closing this in favor of tracking release in #824