qe-wasm: Remove Buffer usage

[SevInf]

Convert Byte arguments to Uint8Array instead.

Allows us to stop polyfilling Buffer for _bg.js fiels in engine.

[github-actions]

WASM Size

Engine	This PR	Base branch	Diff
Postgres	2.043MiB	2.043MiB	-227.000B
Postgres (gzip)	808.896KiB	809.740KiB	-864.000B
Mysql	2.024MiB	2.023MiB	923.000B
Mysql (gzip)	799.862KiB	801.262KiB	-1.401KiB
Sqlite	1.985MiB	1.985MiB	379.000B
Sqlite (gzip)	787.206KiB	787.888KiB	-699.000B

[codspeed-hq]

CodSpeed Performance Report

Merging #4710 will not alter performance

_{Comparing no-buffer (0dd0861) with main (45a0eb4)}

Summary

✅ 11 untouched benchmarks

[github-actions]

❌ WASM query-engine performance will worsen by 1.54%

Full benchmark report

DATABASE_URL="postgresql://postgres:postgres@localhost:5432/bench?schema=imdb_bench&sslmode=disable" \
node --experimental-wasm-modules query-engine/driver-adapters/executor/dist/bench.mjs
cpu: AMD EPYC 7763 64-Core Processor
runtime: node v18.19.0 (x64-linux)

benchmark                   time (avg)             (min … max)       p75       p99      p999
-------------------------------------------------------------- -----------------------------
• movies.findMany() (all - 25000)
-------------------------------------------------------------- -----------------------------
Web Assembly: Baseline  304.49 ms/iter (302.75 ms … 310.16 ms) 305.34 ms 310.16 ms 310.16 ms
Web Assembly: Latest    381.68 ms/iter (378.63 ms … 395.89 ms) 380.74 ms 395.89 ms 395.89 ms
Web Assembly: Current   387.16 ms/iter (384.45 ms … 398.23 ms) 388.65 ms 398.23 ms 398.23 ms
Node API: Current       229.19 ms/iter (225.65 ms … 235.24 ms) 232.73 ms 235.24 ms 235.24 ms

summary for movies.findMany() (all - 25000)
  Web Assembly: Current
   1.69x slower than Node API: Current
   1.27x slower than Web Assembly: Baseline
   1.01x slower than Web Assembly: Latest

• movies.findMany({ take: 2000 })
-------------------------------------------------------------- -----------------------------
Web Assembly: Baseline   12.01 ms/iter   (11.82 ms … 13.69 ms)  11.94 ms  13.69 ms  13.69 ms
Web Assembly: Latest     16.01 ms/iter   (15.31 ms … 20.91 ms)  15.86 ms  20.91 ms  20.91 ms
Web Assembly: Current       16 ms/iter   (15.52 ms … 18.24 ms)  16.07 ms  18.24 ms  18.24 ms
Node API: Current        8,769 µs/iter   (8,531 µs … 9,315 µs)  8,864 µs  9,315 µs  9,315 µs

summary for movies.findMany({ take: 2000 })
  Web Assembly: Current
   1.82x slower than Node API: Current
   1.33x slower than Web Assembly: Baseline
   1x faster than Web Assembly: Latest

• movies.findMany({ where: {...}, take: 2000 })
-------------------------------------------------------------- -----------------------------
Web Assembly: Baseline   1,927 µs/iter   (1,822 µs … 3,347 µs)  1,901 µs  2,848 µs  3,347 µs
Web Assembly: Latest     2,469 µs/iter   (2,384 µs … 3,077 µs)  2,460 µs  3,041 µs  3,077 µs
Web Assembly: Current    2,524 µs/iter   (2,420 µs … 3,711 µs)  2,512 µs  3,604 µs  3,711 µs
Node API: Current        1,490 µs/iter   (1,414 µs … 1,947 µs)  1,501 µs  1,696 µs  1,947 µs

summary for movies.findMany({ where: {...}, take: 2000 })
  Web Assembly: Current
   1.69x slower than Node API: Current
   1.31x slower than Web Assembly: Baseline
   1.02x slower than Web Assembly: Latest

• movies.findMany({ include: { cast: true } take: 2000 }) (m2m)
-------------------------------------------------------------- -----------------------------
Web Assembly: Baseline   12.71 ms/iter   (12.16 ms … 18.43 ms)  12.35 ms  18.43 ms  18.43 ms
Web Assembly: Latest     15.71 ms/iter   (15.61 ms … 15.91 ms)  15.73 ms  15.91 ms  15.91 ms
Web Assembly: Current    16.18 ms/iter   (15.78 ms … 19.95 ms)  15.98 ms  19.95 ms  19.95 ms
Node API: Current        8,866 µs/iter   (8,626 µs … 9,229 µs)  8,929 µs  9,229 µs  9,229 µs

summary for movies.findMany({ include: { cast: true } take: 2000 }) (m2m)
  Web Assembly: Current
   1.83x slower than Node API: Current
   1.27x slower than Web Assembly: Baseline
   1.03x slower than Web Assembly: Latest

• movies.findMany({ where: {...}, include: { cast: true } take: 2000 }) (m2m)
-------------------------------------------------------------- -----------------------------
Web Assembly: Baseline   1,909 µs/iter   (1,829 µs … 2,619 µs)  1,903 µs  2,475 µs  2,619 µs
Web Assembly: Latest     2,469 µs/iter   (2,386 µs … 3,139 µs)  2,460 µs  2,967 µs  3,139 µs
Web Assembly: Current    2,524 µs/iter   (2,419 µs … 3,977 µs)  2,508 µs  3,447 µs  3,977 µs
Node API: Current        1,516 µs/iter   (1,380 µs … 2,396 µs)  1,502 µs  2,207 µs  2,396 µs

summary for movies.findMany({ where: {...}, include: { cast: true } take: 2000 }) (m2m)
  Web Assembly: Current
   1.67x slower than Node API: Current
   1.32x slower than Web Assembly: Baseline
   1.02x slower than Web Assembly: Latest

• movies.findMany({ take: 2000, include: { cast: { include: { person: true } } } })
-------------------------------------------------------------- -----------------------------
Web Assembly: Baseline   12.15 ms/iter   (12.06 ms … 12.33 ms)  12.17 ms  12.33 ms  12.33 ms
Web Assembly: Latest     15.84 ms/iter   (15.54 ms … 20.74 ms)  15.72 ms  20.74 ms  20.74 ms
Web Assembly: Current    15.84 ms/iter   (15.69 ms … 16.35 ms)  15.87 ms  16.35 ms  16.35 ms
Node API: Current        8,814 µs/iter   (8,625 µs … 9,155 µs)  8,876 µs  9,155 µs  9,155 µs

summary for movies.findMany({ take: 2000, include: { cast: { include: { person: true } } } })
  Web Assembly: Current
   1.8x slower than Node API: Current
   1.3x slower than Web Assembly: Baseline
   1x faster than Web Assembly: Latest

• movie.findMany({ where: { ... }, take: 2000, include: { cast: { include: { person: true } } } })
-------------------------------------------------------------- -----------------------------
Web Assembly: Baseline   1,982 µs/iter   (1,823 µs … 3,582 µs)  1,929 µs  3,533 µs  3,582 µs
Web Assembly: Latest     2,419 µs/iter   (2,357 µs … 2,869 µs)  2,426 µs  2,752 µs  2,869 µs
Web Assembly: Current    2,457 µs/iter   (2,396 µs … 2,875 µs)  2,458 µs  2,798 µs  2,875 µs
Node API: Current        1,482 µs/iter   (1,412 µs … 1,960 µs)  1,496 µs  1,687 µs  1,960 µs

summary for movie.findMany({ where: { ... }, take: 2000, include: { cast: { include: { person: true } } } })
  Web Assembly: Current
   1.66x slower than Node API: Current
   1.24x slower than Web Assembly: Baseline
   1.02x slower than Web Assembly: Latest

• movie.findMany({ where: { reviews: { author: { ... } }, take: 100 }) (to-many -> to-one)
-------------------------------------------------------------- -----------------------------
Web Assembly: Baseline  898.37 µs/iter  (846.52 µs … 1,444 µs) 896.23 µs  1,346 µs  1,444 µs
Web Assembly: Latest     1,188 µs/iter   (1,139 µs … 1,846 µs)  1,190 µs  1,465 µs  1,846 µs
Web Assembly: Current    1,230 µs/iter   (1,160 µs … 1,966 µs)  1,225 µs  1,870 µs  1,966 µs
Node API: Current       809.82 µs/iter  (742.43 µs … 1,024 µs)  823.7 µs 960.49 µs  1,024 µs

summary for movie.findMany({ where: { reviews: { author: { ... } }, take: 100 }) (to-many -> to-one)
  Web Assembly: Current
   1.52x slower than Node API: Current
   1.37x slower than Web Assembly: Baseline
   1.04x slower than Web Assembly: Latest

• movie.findMany({ where: { cast: { person: { ... } }, take: 100 }) (m2m -> to-one)
-------------------------------------------------------------- -----------------------------
Web Assembly: Baseline  896.28 µs/iter   (854.7 µs … 1,326 µs) 900.25 µs  1,179 µs  1,326 µs
Web Assembly: Latest     1,200 µs/iter   (1,138 µs … 1,873 µs)  1,196 µs  1,740 µs  1,873 µs
Web Assembly: Current    1,203 µs/iter   (1,154 µs … 1,516 µs)  1,209 µs  1,475 µs  1,516 µs
Node API: Current       795.18 µs/iter   (746.8 µs … 1,145 µs)    819 µs 882.06 µs  1,145 µs

summary for movie.findMany({ where: { cast: { person: { ... } }, take: 100 }) (m2m -> to-one)
  Web Assembly: Current
   1.51x slower than Node API: Current
   1.34x slower than Web Assembly: Baseline
   1x faster than Web Assembly: Latest

After changes in 0dd0861

[Jolg42]

Looks good, though there are some test failures. (I didn't check if it was related or not)

[millsp]

I wondered if we built with wasm-bindgen --browser instead of a Node.js target if that would actually fix the issue in a different way? Since we are targeting browser environments, I guess we should change the current target even if this PR fixes this?

[SevInf]

I wondered if we built with wasm-bindgen --browser instead of a Node.js target if that would actually fix the issue in a different way? Since we are targeting browser environments, I guess we should change the current target even if this PR fixes this?

Buffer usage is on us here, it's not automatically inserted by wasm-bindgen. We used it to convert Bytes arguments for JS consumption