Previously the receiver text was taken directly from the AST, which in macros is missing trivia, leading to corruption (or just unintended replacement of user code).
Now we upmap the range, and extract the original file text in it.
It should be considered by the edition of the caller, not the callee.
Technically we still don't do it correctly - we need the span of the method name (if it comes from a macro), but we don't keep it and this is good enough for now.
fix: In completion's expand, consider recursion stop condition (when we're not inside a macro call anymore) *after* the recursive call instead of before it
This is because our detection is imperfect, and miss some cases such as an impersonating `test` macro, so we hope we'll expand successfully in this case.
Split manual.adoc into markdown files, one for each chapter.
For the parts of the manual that are generated from source code doc
comments, update the comments to use markdown syntax and update the
code generators to write to `generated.md` files.
For the weekly release, stop copying the .adoc files to the
`rust-analyzer/rust-analyzer.github.io` at release time. Instead,
we'll sync the manual hourly from this repository.
See https://github.com/rust-analyzer/rust-analyzer.github.io/pull/226
for the sync. This PR should be merged first, and that PR needs to be
merged before the next weekly release.
This change is based on #15795, but rebased and updated. I've also
manually checked each page for markdown syntax issues and fixed any I
encountered.
Co-authored-by: Lukas Wirth <lukastw97@gmail.com>
Co-authored-by: Josh Rotenberg <joshrotenberg@gmail.com>
This time, when completing the keyword (e.g. `fn` + whitespace).
The bug was actually a double-bug:
First, we did not resolve the impl in the macro-expanded file but in the real file, which of course cannot work.
Second, in analysis the whitespace was correlated with the `impl` and not the incomplete `fn`, which caused fake (where we insert an identifier after the whitespace) and real expansions to go out of sync, which failed analysis. The fix is to skip whitespaces in analysis.
It should be left biased, not right biased, because when e.g. the use has typed `h` then requested completion, the `h` is what we want to find, not the next token (which might indeed be inside a macro call).
I'm not sure why I wrote `right_biased()` to begin with (I remember I had a reason and not just "both should work"), I might've copied the code in `expand_and_analyze()` (which is wrong, because there it lookups on the speculative file, where right biased will always find the correct token and left biased not).
This is still not perfect, because there might not be an identifier already typed then we might still end up in a macro call, but this is the best we can do.
