IDE diagnostic engine currently decides whether or not to run analyzers on closed files based on whether or not the analyzer reports any non-hidden diagnostics. This seems like a very fragile approach, which has caused multiple performance regressions for IDE full solution analysis. Additionally, when an end user bumps up the severity of any IDE hidden diagnostic, they have absolutely no idea that they are paying a very large IDE performance cost.
We should remove this hack in the IDE diagnostic engine and allow the end-user to explicitly configure whether or not to run IDE analyzers for closed files – either per rule basis or a single workspace wide setting. We should extend the IDE tools option UI that allows changing the per-rule severity to add this closed file setting(s). I’ll file an issue to track this.
IDE diagnostic engine currently decides whether or not to run analyzers on closed files based on whether or not the analyzer reports any non-hidden diagnostics. This seems like a very fragile approach, which has caused multiple performance regressions for IDE full solution analysis. Additionally, when an end user bumps up the severity of any IDE hidden diagnostic, they have absolutely no idea that they are paying a very large IDE performance cost.
We should remove this hack in the IDE diagnostic engine and allow the end-user to explicitly configure whether or not to run IDE analyzers for closed files – either per rule basis or a single workspace wide setting. We should extend the IDE tools option UI that allows changing the per-rule severity to add this closed file setting(s). I’ll file an issue to track this.