mirror of
https://gh.wpcy.net/https://github.com/discourse/discourse.git
synced 2026-05-06 21:30:25 +08:00
d446cc7318
Scheduled problem checks with multiple targets are not honouring the `run_every` configuration. For checks with multiple targets, all targets are checked in a single instance of the problem check. However, we have one problem check tracker per target. This mismatch results in the `#ready_to_run?` method always creating a tracker with no target when being checked. This commit fixes that by: **Expect checks to operate on a single target.** This change makes it so that instances of a `ProblemCheck` class are initialized with a target. So instead of 1-N we now have an N-N relationship between checks and trackers. Each instance can access their `target` through an attribute of the same name. This also means problem checks are back to returning a singular `Problem` or `nil`, instead of `[Problem]` or `[]`. For scheduled checks, this means that `ScheduleProblemChecks` now enqueues `N` jobs (where `N` is the number of targets) per check instead of `1` job per check. **Update existing targeted checks to operate on a single target.** This is essentially just removing the loop inside the check.
11 lines
185 B
Ruby
11 lines
185 B
Ruby
# frozen_string_literal: true
|
|
|
|
class ProblemCheck::ForceHttps < ProblemCheck
|
|
self.priority = "low"
|
|
|
|
def call
|
|
return no_problem if SiteSetting.force_https
|
|
|
|
problem
|
|
end
|
|
end
|
