- Don't log entire LSP notifications/requests for the `info` level, instead log of the form:
- `Notification<method>` e.g. Notification<textDocument/publishDiagnostics>
- `Request<method(id)>` e.g. Request<textDocument/hover(6)>
- `Response<method(id)` e.g. Response<textDocument/hover(6)>
- Only log sourcekitd requests/responses at the debug level
This allows building without dispatch and Foundation build roots, which
allows for a multi-phase build. This is intended to support the Windows
staged build where dispatch and Foundation are part of the SDK and
SourceKit-LSP is part of the devtools.
- Avoid short queries (< 3 characters) since they are slow and noisy. Xcode does the same thing.
- Limit the number of results returned:
- Improves performance for queries which match many symbols
- Currently take the first N results to improve performance for large projects
but once we support cancellation we should take the top N.
With this change, tests pass whether we default to server-side or
client-side filtering. Also duplicate a few interesting tests to run
both ways. A future commit will beef up the test coverage for
server-side filtering specifically.
Work in progress: add a basic implementation for keeping track of the
current completion session, and use the sourcekitd
complete.open/update/close requests.
* Forward definition requests when index lookup fails
If we're unable to look up a definition or declaration via the
index, forward the request to the language service.
This allows clangd to support go to definition or declaration
for symbols defined/declared in the AST.
While fallback arguments are being used (either from the fallback build system or fallback settings from the primary build system), withhold semantic diagnostics from sourcekitd and all diagnostics from clangd. This helps prevent user confusion from spurious errors.
- Also remove the DocumentURI standardization in favor of proper equality + hash checks to work around the %40 --> @ encoding issue seen on CI.
