perf: optimize OperatorsReader::skip_const_expr

[dishmaker]

This optimizes just OperatorsReader::skip_const_expr from:

• perf: remove pointer-dereference using owned version of visit_operator #2638

bench parse/tests change: -2%

[Spoiler] cargo bench

ran 4 times (main and my branch)

parse/tests             time:   [7.5023 ms 7.5062 ms 7.5105 ms]
                        change: [−2.2118% −2.0553% −1.9033%] (p = 0.00 < 0.05)
                        Performance has improved.
parse/tests             time:   [7.4871 ms 7.4903 ms 7.4937 ms]
                        change: [−2.5212% −2.4240% −2.3359%] (p = 0.00 < 0.05)
                        Performance has improved.
parse/tests             time:   [7.4769 ms 7.4807 ms 7.4849 ms]
                        change: [−2.4526% −2.3487% −2.2449%] (p = 0.00 < 0.05)
                        Performance has improved.                
parse/tests             time:   [7.4658 ms 7.4712 ms 7.4768 ms]
                        change: [−2.0683% −1.9672% −1.8664%] (p = 0.00 < 0.05)
                        Performance has improved.


                        
validate/tests          time:   [43.889 ms 43.930 ms 43.977 ms]
                        change: [+0.9219% +1.0458% +1.1819%] (p = 0.00 < 0.05)
                        Change within noise threshold.
validate/tests          time:   [43.413 ms 43.440 ms 43.467 ms]
                        change: [−0.5937% −0.4812% −0.3684%] (p = 0.00 < 0.05)
                        Change within noise threshold.
validate/tests          time:   [43.516 ms 43.570 ms 43.634 ms]
                        change: [−0.0868% +0.0767% +0.2552%] (p = 0.36 > 0.05)
                        No change in performance detected.
validate/tests          time:   [43.715 ms 43.747 ms 43.782 ms]
                        change: [−0.0765% +0.0745% +0.2170%] (p = 0.32 > 0.05)
                        No change in performance detected.

[alexcrichton]

Thanks!