This binary search implementation differs from Rust's
`slice::binary_search_by` method in how they deal with ties.
In Rust's implementation:
> If there are multiple matches, then any one of the matches
> could be returned.
This implementation needs to return the index of the *first* match.
This PR adds an `--update` flag to the `tree-sitter test` command, which adds the ability to replace the _expected_ output in the corpus.txt with the _actual_ output produced by the parser, that is, we can now simply use this `--update` flag to write all the corresponding parser output back to the corpus.txt, and we just need to check the output without typing its actual sexp.
- use the same output format as `tree-sitter parse`, except there won't be any position information printed.
- the corpus.txt won't be touched if there's no difference between the _expected_ output and the _actual_ output in that file.
- if there're differences between _expected_ and _actual_, _expected_ will be replaced by _actual_ and the whole file will be reformatted, i.e., all the output sexp will be formatted just like the output from `tree-sitter parse` and all the delimiters `===`/`---` will be normalized as 80-column long.
- this flag also works with `--filter` flag.
When debugging a test with 'script/test -D', the DOT-graph generation
code was sometimes causing reallocations that were not captured by the
allocation tracker, because we explicitly disable allocation-tracking
for that method in order to reduce noise when debugging memory leaks.
By growing the relevant array *prior* to turning off allocation
tracking, we can ensure that it is not reallocated within that function,
avoiding false positive memory leak errors.
Fixes#302
When there are embedded documents, multiple scopes can start or
end at the same position. Previously, there was no guarantee that
the ScopeEnd events would always occur in the reverse order of the
ScopeStart events. The easiest way to avoid exposing inconsistency
is to not surface the scopes being ended.
