-
Notifications
You must be signed in to change notification settings - Fork 1.7k
Request: ability to configure the analysis server #23851
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
Comments
Initial thoughts:
|
Here's a sample of the kind of format we've been discussing:
Linter config ideas derived from #57189 and included just for context. Consuming linter options will be the province of the linter. The most salient bit is the idea that it's YAML and has a mapping from 'scopes' to options. For further reference, here's where we parse options in And here's where the options format is defined for plugins: Regarding discovery by server, I think we should allow options to live at all "context roots" (e.g., next to a |
Maybe the file should be read by |
Longer term we plan on allowing But as it stands today, I don't think it's useful to run Just to be clear, this won't prevent analyzer from analyzing a file that is referenced from a non-ignored file. The semantics are more accurately described as: don't analyze any file that matches one of the given patterns unless it's necessary in order to analyze a file that doesn't match one of the given files. |
By the way, the previous comment would have been easier to write if the second-level key ('ignore') were instead 'exclude' (because 'exclude' has a nice opposite -- 'include' -- while 'ignore' doesn't, unless it's something like 'attend', but 'attended files' just doesn't sound right). |
sgtm (also, bumping to High because it's really important for our Sky customers, and we're starting on it anyway :) |
With a CLI option it already sort of is (though no-one is doing anything with it). Interested plugins are all passed the contents for independent processing. Migrating the linter to use this mechanism is on the short-list. (And needless to say none of this include/exclude configuration logic is implemented yet either.) As Brian says, discovery will come down the road.
👍 More sensibly you would want lint (and other plugin) configurations to get picked up though I think.
I noticed that as I was putting together this proposal. If we also want an opposite (and I don't see why we should rule it out), let's go with Any objections? |
Initial support is done. Analysis Service now reads a .analysis_config YAML file. |
We'd like to be able to configure the analysis server, in such as way that the configuration is associated/tied to a project.
There are a few things we'd like to configure, among other things:
Ideally, this configuration is human readable/writable, is machine readable/writable, is easy to extend, can be checked into source control, and is discoverable by the analysis server.
The text was updated successfully, but these errors were encountered: