expression, planner: support builtin function benchmark

[wuudjac]

What problem does this PR solve?

Implements builtin function BENCHMARK(), and introduces a counter in expression_rewriter to disable constant folding. (#6774)

What is changed and how it works?

1. Implements builtin BENCHMARK() function.
2. To disable constant folding inside BENCHMARK() scope, expressionRewriter.disableFoldCounter is introduced.

Check List

Tests

• Unit test
• Integration test
• Manual test

select benchmark(1000000000, 3);
select benchmark(1000000000, length("abc"));
select benchmark(1000000000, sin(123) + sin(456));

Execurtion time of these 3 commands should differ.

Code changes

• builtinBenchmarkSig in expression/builtin_info.go
• expressionRewriter.disableFoldCounter in planner/core/expression_rewriter.go
• expressionRewriter.newFunction() in planner/core/expression_rewriter.go to control new function behavior by the counter.

Side effects

• Everytime the expression_rewriter wants to get a new builtin function expression, it has to use expressionRewriter.newFunction() instead of expression.NewFunction(), or this builtin function may not be available to be measured by BENCHMARK() because of constant folding.

Related changes

• Nothing.

[CLAassistant]

All committers have signed the CLA.

[codecov-io]

Codecov Report

Merging #9252 into master will increase coverage by <.01%.
The diff coverage is 72.22%.

@@            Coverage Diff             @@
##           master    #9252      +/-   ##
==========================================
+ Coverage   67.15%   67.15%   +<.01%     
==========================================
  Files         371      371              
  Lines       77427    77491      +64     
==========================================
+ Hits        51995    52041      +46     
- Misses      20785    20797      +12     
- Partials     4647     4653       +6

Impacted Files	Coverage Δ
expression/builtin_info.go	61.94% <64.91%> (+0.77%)	⬆️
planner/core/expression_rewriter.go	73.93% <84.84%> (+0.31%)	⬆️
store/tikv/scan.go	73.94% <0%> (-3.37%)	⬇️
infoschema/infoschema.go	76.31% <0%> (-1.32%)	⬇️
expression/schema.go	93.75% <0%> (-0.79%)	⬇️
executor/join.go	77.86% <0%> (-0.53%)	⬇️
expression/builtin_string.go	76% <0%> (+0.1%)	⬆️
expression/util.go	72.57% <0%> (+0.57%)	⬆️
expression/scalar_function.go	92.64% <0%> (+2.2%)	⬆️
util/systimemon/systime_mon.go	100% <0%> (+20%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 357f9d7...555686b. Read the comment docs.

[shenli]

@wuudjac Thanks for your contribution!

[zz-jason]

LGTM

[alivxxx]

LGTM

[alivxxx]

/run-all-tests