[DavidBadura]

Follow-up to the concurrency groups added in the last pull request. They used the workflow name plus the branch name, which lets two workflows with the same name cancel each other on the same branch, and two fork pull requests with the same branch name (e.g. patch-1) cancel each other. github.workflow_ref is the workflow file plus the ref (refs/pull/<number>/merge for pull requests) and has neither problem.

[github-actions]

Hello 👋

Here are the most recent benchmark results, compared against 2.1.x. Time is the mode over the iterations in the its column, with the relative standard deviation in brackets.

StackHydrator

subject	revs	its	memory	base	current	diff
benchHydrate1Object	1000	5	1.855mb -21.23%	1.863μs (±1.92%)	1.898μs (±1.15%)	+1.85%
benchExtract1Object	1000	5	1.855mb -21.23%	2.145μs (±1.08%)	2.151μs (±0.67%)	+0.28%
benchHydrate1000Objects	3	5	1.855mb 0.00%	1.894ms (±0.92%)	1.883ms (±0.67%)	-0.62%
benchExtract1000Objects	3	5	1.855mb 0.00%	1.768ms (±0.84%)	1.771ms (±1.19%)	+0.17%

StackHydratorWithLazy

subject	revs	its	memory	base	current	diff
benchHydrate1Object	1000	5	1.855mb 0.00%	0.288μs (±2.08%)	0.292μs (±4.22%)	+1.52%
benchHydrate1ObjectTriggerInit	1000	5	1.855mb 0.00%	2.234μs (±1.05%)	2.233μs (±1.23%)	-0.05%
benchHydrate1000Objects	3	5	1.855mb 0.00%	282.452μs (±2.22%)	281.624μs (±2.50%)	-0.29%
benchHydrate1000ObjectsTriggerInit	3	5	1.855mb 0.00%	2.256ms (±0.49%)	2.246ms (±0.60%)	-0.44%

StackHydratorWithCryptography

subject	revs	its	memory	base	current	diff
benchHydrate1Object	1000	5	1.855mb 0.00%	2.780μs (±0.51%)	2.736μs (±1.91%)	-1.60%
benchExtract1Object	1000	5	1.855mb 0.00%	6.759μs (±0.94%)	6.840μs (±1.83%)	+1.19%
benchHydrate1000Objects	3	5	1.855mb 0.00%	2.726ms (±0.64%)	2.726ms (±0.52%)	+0.00%
benchExtract1000Objects	3	5	1.855mb 0.00%	6.409ms (±0.82%)	6.408ms (±0.49%)	-0.02%

This comment gets updated every time a new commit comes in.