Upon a `textDocument/semanticTokens/full` request the analysis server
sorts the "tokens" via `sort` on List giving it a compare method.
The method, though, created a map literal for each comparison.
This CL "inlines" the map (it only had one entry anyway) and adds a
benchmark, where I note these changes:
```
1000:
Difference at 95.0% confidence
-0.0274151 +/- 0.0135978
-33.1981% +/- 16.4662%
(Student's t, pooled s = 0.0181832)
2000:
Difference at 95.0% confidence
-0.027344 +/- 0.0096196
-20.9529% +/- 7.3712%
(Student's t, pooled s = 0.0128635)
4000:
Difference at 95.0% confidence
-0.108847 +/- 0.0196627
-35.3172% +/- 6.37987%
(Student's t, pooled s = 0.0262932)
8000:
Difference at 95.0% confidence
-0.20647 +/- 0.0272577
-34.4932% +/- 4.55371%
(Student's t, pooled s = 0.0364494)
16000:
Difference at 95.0% confidence
-0.384931 +/- 0.0401393
-31.4809% +/- 3.28272%
(Student's t, pooled s = 0.0536748)
```
Change-Id: Ibacce056cd3f154afc9377805bbccfd8c5ca9ea9
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/456341
Reviewed-by: Brian Wilkerson <brianwilkerson@google.com>
Commit-Queue: Jens Johansen <jensj@google.com>
Jens Johansen
820a1ec5a5