examples(allocation): ensures data used in host functions isn't freed before invocation

[lburgazzoli]

No description provided.

[codefromthecrypt]

Thanks! mind doing this and sharing (before/after bench) in the PR description?

# from main
go test -run='^$' -bench '^BenchmarkAllocation.*' ./internal/integration_test/vs/compiler -count=6 > old.txt

# from this commit
go test -run='^$' -bench '^BenchmarkAllocation.*' ./internal/integration_test/vs/compiler -count=6 > new.txt

benchstat old.txt new.txt 

[lburgazzoli]

➜ benchstat old.txt new.txt 


goos: linux
goarch: amd64
pkg: github.com/tetratelabs/wazero/internal/integration_test/vs/compiler
cpu: Intel(R) Core(TM) i7-10850H CPU @ 2.70GHz
                          │   old.txt    │              new.txt               │
                          │    sec/op    │    sec/op     vs base              │
Allocation/Compile-12       15.63m ± 61%   15.57m ± 82%       ~ (p=0.937 n=6)
Allocation/Instantiate-12   353.2µ ± 17%   357.4µ ±  5%       ~ (p=0.937 n=6)
Allocation/Call-12          3.212µ ±  2%   2.898µ ±  2%  -9.78% (p=0.002 n=6)
geomean                     260.8µ         252.6µ        -3.11%

                          │   old.txt    │               new.txt                │
                          │     B/op     │     B/op      vs base                │
Allocation/Compile-12       2.817Mi ± 0%   2.816Mi ± 0%       ~ (p=0.065 n=6)
Allocation/Instantiate-12   351.4Ki ± 0%   351.4Ki ± 0%       ~ (p=0.818 n=6)
Allocation/Call-12            48.00 ± 0%     48.00 ± 0%       ~ (p=1.000 n=6) ¹
geomean                     36.22Ki        36.21Ki       -0.01%
¹ all samples are equal

                          │   old.txt   │               new.txt               │
                          │  allocs/op  │  allocs/op   vs base                │
Allocation/Compile-12       2.022k ± 0%   2.012k ± 0%  -0.49% (p=0.002 n=6)
Allocation/Instantiate-12    852.0 ± 0%    852.0 ± 0%       ~ (p=1.000 n=6) ¹
Allocation/Call-12           5.000 ± 0%    5.000 ± 0%       ~ (p=1.000 n=6) ¹
geomean                      205.0         204.7       -0.17%
¹ all samples are equal

[codefromthecrypt]

code wise I like it.

While personally, I rarely understand when I'm supposed to do runtime.KeepAlive, this seems the pattern to do it. What I like also is that the stringToPtr code is more concise.

One homework before merge is checking that unsafe.StringData won't return nil when compiled with TinyGo's wasi target.

[ncruces]

While we're at it, is there a reason not to do the below change:

wazero/examples/allocation/tinygo/testdata/greet.go

Lines 66 to 68 in 3c01f86

	func ptrToString(ptr uint32, size uint32) string {
	return unsafe.String((*byte)(unsafe.Pointer(uintptr(ptr))), size)
	}

Seems to match guidance from tinygo-org/tinygo#1284

[codefromthecrypt]

Sorry I wasn't sure which line we were proceeding with, so I commented on #1436 (review) instead of this one. Feel free to proceed forward with whatever, as I've already bookmarked 1436 for next time

[ncruces]

I think we should merge this one. I did my changes offline and didn't notice this PR before push.

[lburgazzoli]

No objection merge either this or #1436 , they are essentially equivalent

[evacchi]

merged #1436 thanks @lburgazzoli