Stephan Renatus 839ea94f12 mask: don't rely on JSON roundtrip to convert structs (#5770) ...

Reading the code, I couldn't bear the use of a JSON roundtrip to convert
from one kind of struct to another.

$ benchstat main.out pr.out
goos: darwin
goarch: amd64
pkg: github.com/open-policy-agent/opa/plugins/logs
cpu: Intel(R) Core(TM) i9-9880H CPU @ 2.30GHz
                                  │  main.out   │               pr.out                │
                                  │   sec/op    │   sec/op     vs base                │
MaskingNop-16                       4.779µ ± 2%   4.892µ ± 1%   +2.38% (p=0.009 n=10)
MaskingRuleCountsNop/1Rules-16      4.859µ ± 3%   4.868µ ± 1%        ~ (p=0.754 n=10)
MaskingRuleCountsNop/10Rules-16     5.047µ ± 3%   4.872µ ± 0%   -3.47% (p=0.000 n=10)
MaskingRuleCountsNop/100Rules-16    5.089µ ± 4%   4.846µ ± 1%   -4.78% (p=0.000 n=10)
MaskingRuleCountsNop/1000Rules-16   4.894µ ± 4%   4.905µ ± 3%        ~ (p=0.684 n=10)
MaskingErase-16                     14.66µ ± 2%   11.77µ ± 3%  -19.72% (p=0.000 n=10)
geomean                             5.914µ        5.648µ        -4.50%

                                  │   main.out   │                 pr.out                 │
                                  │     B/op     │     B/op      vs base                  │
MaskingNop-16                       3.136Ki ± 0%   3.136Ki ± 0%        ~ (p=1.000 n=10)
MaskingRuleCountsNop/1Rules-16      3.136Ki ± 0%   3.136Ki ± 0%        ~ (p=1.000 n=10) ¹
MaskingRuleCountsNop/10Rules-16     3.136Ki ± 0%   3.136Ki ± 0%        ~ (p=1.000 n=10) ¹
MaskingRuleCountsNop/100Rules-16    3.136Ki ± 0%   3.136Ki ± 0%        ~ (p=1.000 n=10) ¹
MaskingRuleCountsNop/1000Rules-16   3.136Ki ± 0%   3.136Ki ± 0%        ~ (p=1.000 n=10) ¹
MaskingErase-16                     8.328Ki ± 0%   7.281Ki ± 0%  -12.57% (p=0.000 n=10)
geomean                             3.690Ki        3.608Ki        -2.21%
¹ all samples are equal

                                  │  main.out  │               pr.out                │
                                  │ allocs/op  │ allocs/op   vs base                 │
MaskingNop-16                       65.00 ± 0%   65.00 ± 0%       ~ (p=1.000 n=10) ¹
MaskingRuleCountsNop/1Rules-16      65.00 ± 0%   65.00 ± 0%       ~ (p=1.000 n=10) ¹
MaskingRuleCountsNop/10Rules-16     65.00 ± 0%   65.00 ± 0%       ~ (p=1.000 n=10) ¹
MaskingRuleCountsNop/100Rules-16    65.00 ± 0%   65.00 ± 0%       ~ (p=1.000 n=10) ¹
MaskingRuleCountsNop/1000Rules-16   65.00 ± 0%   65.00 ± 0%       ~ (p=1.000 n=10) ¹
MaskingErase-16                     149.0 ± 0%   138.0 ± 0%  -7.38% (p=0.000 n=10)
geomean                             74.64        73.69       -1.27%
¹ all samples are equal

Signed-off-by: Stephan Renatus <stephan@styra.com>

2023-03-18 07:55:40 +01:00

..

status

plugins: Surface decision log errors via status API

2023-03-08 00:12:42 -08:00

buffer_test.go

Reduce memory overhead of decision logs

2018-06-15 09:23:01 -07:00

buffer.go

Remove decision log buffer limit

2018-06-18 11:32:22 -07:00

encoder_test.go

Improve error message when chunk size > limit

2022-09-19 18:20:19 -04:00

encoder.go

linters: add unconvert (#5318)

2022-10-27 13:35:39 +02:00

mask_test.go

mask: don't rely on JSON roundtrip to convert structs (#5770)

2023-03-18 07:55:40 +01:00

mask.go

mask: don't rely on JSON roundtrip to convert structs (#5770)

2023-03-18 07:55:40 +01:00

plugin_benchmark_test.go

plugins/logs: don't convert EventV1 to ast.Value twice (#5761)

2023-03-17 10:31:43 +01:00

plugin_test.go

plugins/logs: don't convert EventV1 to ast.Value twice (#5761)

2023-03-17 10:31:43 +01:00

plugin.go

plugins/logs: don't convert EventV1 to ast.Value twice (#5761)

2023-03-17 10:31:43 +01:00