Force warnings on lints #434
Labels
major-change
A proposal to make a major change to rustc
major-change-accepted
A major change proposal that was accepted
T-compiler
Add this label so rfcbot knows to poll the compiler team
Proposal
Add mechanism for forcing lints to trigger.
Summary
--force-warn XXX
option that is given either a lint or lint groupXXX
XXX
will always issue warnings, regardless of whether they are "allowed" by the code or command-line optionscargo fix --edition
in order to force migration lints etc to be consideredMotivation
We would like to take some existing warnings and make them into errors in the new edition. We anticipate this being a common pattern. The problem is that these warnings, if they already exist, may be marked a
#[allow]
in various code bases. As a result,cargo fix
would not see the migration suggestions and migration would not succeed.Proposed plan
To become part of a migration, existing lints can be directly added into
rust-2021-migration
group.Caveat: Multiple groups
There has been some concern about members of multiple groups.
This plan may mean that they are members of multiple groups. This is generally discouraged for reasons not clearly documented -- Niko's hypothesis is that the reason for this is that (e.g.) the code doesn't consider intersection between two groups, and hence
#[allow(group1)]
and#[warn(group2)]
may result in eitherallow
orwarn
for any common members, depending on the order in which those directives are processed. Perhaps others can confirm. This seems like an independent bug that should be fixed -- however, using this command line option, it is not a big probem because the members ofrust-2021-migrations
will be warn regardless of what other groups they may belong to.There may be confusion in other parts of the code, such as diagnostics.
Also, we don't expect users to commonly add
rust_2021_migrations
.This plan means that some lints are members of multiple groups. This has been discouraged but in the past but we currently believe that it should work fine. We should test the scenario where a lint is in two groups and one of those groups is allow.
Alternatives
We could instead introduce a fresh copy of these lints that is a member of
rust_2021_migrations
. For example, if there is a lintfoo
, maybe we make afoo_2021
lint that is specifically for the migration. We could perhaps make this convenient to issue in the code by having some option when creating the lint that is like.migration(RUST_2021_MIGRATioN)
. This could also make the lint into a hard error automatically in the new edition, regardless of the lint level.Mentors or Reviewers
@nikomatsakis
Process
The main points of the Major Change Process are as follows:
@rustbot second
.-C flag
, then full team check-off is required.@rfcbot fcp merge
on either the MCP or the PR.You can read more about Major Change Proposals on forge.
Comments
This issue is not meant to be used for technical discussion. There is a Zulip stream for that. Use this issue to leave procedural comments, such as volunteering to review, indicating that you second the proposal (or third, etc), or raising a concern that you would like to be addressed.
The text was updated successfully, but these errors were encountered: