Sync with go1.23

[liggitt]

First commit is a straight copy of the json/encoding files from stdlib on the go 1.23 release branch

Second commit is a reapply of carry patches to those files

Third commit is a tweak to make the external dependency check work with go 1.23 go mod graph behavior

[jpbetz]

I compared the carry patch commit against the "patch:..." commits in the repo history and it all looks right to me. The other changes look good too.

/approve
/lgtm

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: jpbetz, liggitt

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [liggitt]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[liggitt]

oof... sorry... this was still WIP ... it appears the stdlib json byte use may have regressed between 1.19 and 1.23 and I was chasing that down... will follow that up and resolve or revert before bumping in k/k

[liggitt]

benchcmp 1.20.14.txt 1.21.13.txt 
benchmark                                old ns/op     new ns/op     delta
BenchmarkUnmarshal/typed_stdlib-10       33821         32411         -4.17%
BenchmarkUnmarshal/typed_stdlib-10       33654         32313         -3.98%
BenchmarkUnmarshal/typed_stdlib-10       33417         32258         -3.47%
BenchmarkUnmarshal/typed_stdlib-10       33361         32598         -2.29%
BenchmarkUnmarshal/typed_stdlib-10       33344         32292         -3.15%
BenchmarkUnmarshal/typed_stdlib-10       33385         32259         -3.37%
BenchmarkUnmarshal/typed_stdlib-10       33198         32223         -2.94%
BenchmarkUnmarshal/typed_stdlib-10       33390         32296         -3.28%
BenchmarkUnmarshal/typed_stdlib-10       34015         32165         -5.44%
BenchmarkUnmarshal/typed_stdlib-10       34059         33392         -1.96%
BenchmarkUnmarshal/untyped_stdlib-10     39088         36741         -6.00%
BenchmarkUnmarshal/untyped_stdlib-10     38622         35609         -7.80%
BenchmarkUnmarshal/untyped_stdlib-10     38615         37185         -3.70%
BenchmarkUnmarshal/untyped_stdlib-10     41221         35654         -13.51%
BenchmarkUnmarshal/untyped_stdlib-10     38330         35452         -7.51%
BenchmarkUnmarshal/untyped_stdlib-10     38662         35238         -8.86%
BenchmarkUnmarshal/untyped_stdlib-10     38069         35582         -6.53%
BenchmarkUnmarshal/untyped_stdlib-10     38133         36464         -4.38%
BenchmarkUnmarshal/untyped_stdlib-10     37744         35498         -5.95%
BenchmarkUnmarshal/untyped_stdlib-10     38408         35433         -7.75%

benchmark                                old allocs     new allocs     delta
BenchmarkUnmarshal/typed_stdlib-10       161            162            +0.62%
BenchmarkUnmarshal/typed_stdlib-10       161            162            +0.62%
BenchmarkUnmarshal/typed_stdlib-10       161            162            +0.62%
BenchmarkUnmarshal/typed_stdlib-10       161            162            +0.62%
BenchmarkUnmarshal/typed_stdlib-10       161            162            +0.62%
BenchmarkUnmarshal/typed_stdlib-10       161            162            +0.62%
BenchmarkUnmarshal/typed_stdlib-10       161            162            +0.62%
BenchmarkUnmarshal/typed_stdlib-10       161            162            +0.62%
BenchmarkUnmarshal/typed_stdlib-10       161            162            +0.62%
BenchmarkUnmarshal/typed_stdlib-10       161            162            +0.62%
BenchmarkUnmarshal/untyped_stdlib-10     517            515            -0.39%
BenchmarkUnmarshal/untyped_stdlib-10     517            515            -0.39%
BenchmarkUnmarshal/untyped_stdlib-10     517            515            -0.39%
BenchmarkUnmarshal/untyped_stdlib-10     516            515            -0.19%
BenchmarkUnmarshal/untyped_stdlib-10     517            515            -0.39%
BenchmarkUnmarshal/untyped_stdlib-10     517            515            -0.39%
BenchmarkUnmarshal/untyped_stdlib-10     517            515            -0.39%
BenchmarkUnmarshal/untyped_stdlib-10     517            515            -0.39%
BenchmarkUnmarshal/untyped_stdlib-10     517            515            -0.39%
BenchmarkUnmarshal/untyped_stdlib-10     517            515            -0.39%

benchmark                                old bytes     new bytes     delta
BenchmarkUnmarshal/typed_stdlib-10       13240         20384         +53.96%
BenchmarkUnmarshal/typed_stdlib-10       13240         20384         +53.96%
BenchmarkUnmarshal/typed_stdlib-10       13240         20384         +53.96%
BenchmarkUnmarshal/typed_stdlib-10       13240         20384         +53.96%
BenchmarkUnmarshal/typed_stdlib-10       13240         20384         +53.96%
BenchmarkUnmarshal/typed_stdlib-10       13240         20384         +53.96%
BenchmarkUnmarshal/typed_stdlib-10       13240         20384         +53.96%
BenchmarkUnmarshal/typed_stdlib-10       13240         20384         +53.96%
BenchmarkUnmarshal/typed_stdlib-10       13240         20384         +53.96%
BenchmarkUnmarshal/typed_stdlib-10       13240         20384         +53.96%
BenchmarkUnmarshal/untyped_stdlib-10     28163         27733         -1.53%
BenchmarkUnmarshal/untyped_stdlib-10     28165         27737         -1.52%
BenchmarkUnmarshal/untyped_stdlib-10     28167         27733         -1.54%
BenchmarkUnmarshal/untyped_stdlib-10     28159         27734         -1.51%
BenchmarkUnmarshal/untyped_stdlib-10     28165         27734         -1.53%
BenchmarkUnmarshal/untyped_stdlib-10     28165         27736         -1.52%
BenchmarkUnmarshal/untyped_stdlib-10     28161         27735         -1.51%
BenchmarkUnmarshal/untyped_stdlib-10     28167         27734         -1.54%
BenchmarkUnmarshal/untyped_stdlib-10     28160         27739         -1.50%
BenchmarkUnmarshal/untyped_stdlib-10     28161         27736         -1.51%

[liggitt]

yeah, 50% regression in byte use in typed unmarshals from stdlib 1.20 → 1.21

[jpbetz]

oof... sorry... this was still WIP

My bad. Happy to review follow up when it is ready.

[liggitt]

full change from 1.20 to 1.21:

git log --format=oneline upstream/release-branch.go1.20..upstream/release-branch.go1.21 -- src/encoding/json/

• golang/go@04c6289 encoding/json: document unsupported float values
• golang/go@2e51f6f encoding/json: make use of Buffer.AvailableBuffer
• golang/go@b9b8cec encoding/json: simplify folded name logic
• golang/go@2de406b encoding/json: use append for HTMLEscape
• golang/go@cb3e170 encoding/json: rely on reflect.Value.SetZero
• golang/go@8367e2d encoding/json: remove legacy fuzz.go file
• golang/go@ac27b4d encoding/json: rely on reflect.Value.Grow
• golang/go@8e5f56a encoding/json: use append-like operations for encoding
• golang/go@e2f4134 encoding/json: unify encodeState.string and encodeState.stringBytes
• golang/go@21ff670 encoding/json: use append for Compact and Indent

[liggitt]

opened a revert of this at #20 while I dig into the regression

[liggitt]

bisected to golang/go@ac27b4d encoding/json: rely on reflect.Value.Grow

[liggitt]

captured test results at https://gist.github.com/liggitt/bf28159afdaabf52b78ac43d86bc6362, will open upstream go issue