fix: performance by Siolto · Pull Request #658 · ui5-community/wdi5

Siolto · 2025-02-11T23:10:26Z

First performance improvement of the v9 support.

current improvement for the basic.test.js from ~33 sec to ~13 sec. No further performance tests happened yet.

it was a late night session so the description will get updated. But not yet...

Siolto · 2025-02-12T09:51:04Z

I call it the Promisified callback hell 🔥 🔥

As far as I can tell the performance issues are a combination of our own waitForUI5, which is not fully promisified and still uses callbacks and the use of these callbacks in executeControlMethod and executeObjectMethod.

Current solution idea:

DON'T use any callbacks in waitForUI5 => for the time being, use the RecordReplay.waitForUI5
check if we can get rid of the callbacks in our own waitForUI5 function or even if we still need it for performance reasons.
I am not sure if this will solve all our performance issues but at least for me it makes it a lot faster

Ping: @vobu @marianfoo @hmanchev @monavari-lebrecht @nicoschoenteich

…thod

…ased on protocol

…ndling

…rmance

Siolto · 2025-02-12T20:04:30Z

Performance update

I had to slow down BIDI due to fragile tests (Promise.all is deadly 💀). This comparison is not 100% accurate as I had to fix some tests on the way but we can see the improvements.

There are some BIDI related topics we have to discuss (marked with TODO in the PR) before merging this.

old BIDI performance:

new BIDI performance:

old webdriver performance:

new webdriver performance:

vobu · 2025-02-13T18:55:07Z

you clearly are on a spree here and got further than any of us before. Kudos 🙇
Are the old..new comparisons relating to wdi5^2 vs wdi5^3 or is this wdi5^3-rc.0 vs the current codebase?

Siolto · 2025-02-24T18:56:35Z

you clearly are on a spree here and got further than any of us before. Kudos 🙇 Are the old..new comparisons relating to wdi5^2 vs wdi5^3 or is this wdi5^3-rc.0 vs the current codebase?

That was wdi5^3-rc.0 vs current codebase. For a fair comparison wdi5^2 vs current codebase you have to align the number of maximum instances since the default has increased with v3. If you want you can do this missing comparison 😉

My assumption is that we are still around 10-20% slower with the current codebase, but I am optimistic that we can reduce this even more.

Siolto · 2025-04-10T18:37:01Z

Removed our own waitForUI5 and use the RecordReplay.waitForUI5. I think this first improvement is good to go to get some things moving with wdi5^3.x

[skip ci]

vobu · 2025-05-20T13:29:33Z

thanks to the massive work of @Siolto, we're now seeing a ~20% performance increase in this PR 🚤 🥳

subject: 3 testruns of 25 test suites

wdi5^2: (average of) 1:36 min (WebdriverIO 8, webdriver protocol)
wdi5^3: (average of) 1:16 min (WebdriverIO 9, both with webdriver and bidi)

…I5Ready for consistency

vobu · 2025-05-21T12:42:29Z

examples/ui5-js-app/webapp/test/e2e/exec.test.js

+            // compare results
+            regularPeopleListNames.forEach((name) => {
+                expect(peopleListNames).toContain(name)
+            })


@Siolto
browser.execute is aimed at synchronous callbacks
but can handle async ones
however: if Promise.all is used, and all Promises are unwrapped inside out
we have https://github.com/webdriverio/webdriverio/blob/b70285e4599408da13479ebef0fb95a8cb13d60f/packages/webdriverio/src/commands/browser/execute.ts#L61 resolving
the node <- wdi5 -> browser communication prematurely, resulting in "undefined" results

* fix: performance * fix: improve error handling and result processing in executeControlMethod * fix: refactor executeObjectMethod not using callbacks * test: update DateTimePicker API test to use ISO string format * fix: update generated-methods test to handle different date formats based on protocol * fix: await browser.asControl in press test for proper asynchronous handling * fix: bidi race condition in retrieveElements * fix: update Planning Calendar test to handle BIDI-specific DOM controls * fix: optimize control retrieval for BIDI by removing unnecessary await * fix: adjust BIDI handling in exec test to avoid Promise.all for performance * chore: configuration * fix: add TODO comment for handling UI5 injection options * refactor: remove propietary waitForUI5 * refactor: run planningCalendar test also with BiDi protocol * fix: improve peformance while retrieving multiple elements (ui5-community#665) [skip ci] * refactor: rename clientSide__checkForUI5Ready to clientSide_checkForUI5Ready for consistency * chore: lock file * fix: test * refactor: rm unnecessary await --------- Co-authored-by: Volker Buzek <dev@metalodge.com>

* ci: Improve ESLint v9 config * ci: Bump GH Actions checkout and setup-node to v4 * ci: Bump GH Actions github/codeql-action/* to v3 * ci: Bump GH Actions download-artifact and upload-artifact to v4 * ci: Bump prettier version * chore: Bump types versions * chore: Linting + types * fix: Mostly any types * fix: Remove deprecated browser.switchToFrame * fix: use crypto.randomUUID * refactor: load UI5 libs in 1 task * fix: remove multiple any types * chore: linting * fix: performance (#658) * fix: performance * fix: improve error handling and result processing in executeControlMethod * fix: refactor executeObjectMethod not using callbacks * test: update DateTimePicker API test to use ISO string format * fix: update generated-methods test to handle different date formats based on protocol * fix: await browser.asControl in press test for proper asynchronous handling * fix: bidi race condition in retrieveElements * fix: update Planning Calendar test to handle BIDI-specific DOM controls * fix: optimize control retrieval for BIDI by removing unnecessary await * fix: adjust BIDI handling in exec test to avoid Promise.all for performance * chore: configuration * fix: add TODO comment for handling UI5 injection options * refactor: remove propietary waitForUI5 * refactor: run planningCalendar test also with BiDi protocol * fix: improve peformance while retrieving multiple elements (#665) [skip ci] * refactor: rename clientSide__checkForUI5Ready to clientSide_checkForUI5Ready for consistency * chore: lock file * fix: test * refactor: rm unnecessary await --------- Co-authored-by: Volker Buzek <dev@metalodge.com> * fix: use crypto.randomUUID * fix: revert * build: clean script * test: add more UI5 LTS versions * refactor: remove fs-extra * chore: pretty sure package "i" was installed by mistake * test: fix expected value * chore: update workflows to also run against v9-support * test: use a public method to test selected element * test: expect().toHaveTextContaining() doesn't exist * test: replace getInitStatus() by isInitialized() * test: fix wdio-ui5tooling.conf.js * test: remove deepmerge dependency * chore: npm shouldn't be a dependency * test: simplify test deps with peerDeps * test: replace sinon by node:test * fix: remove devtools dep * fix: use browserInstance rather than global browser * fix: no console - use wdi5 Logger * refactor: compare-versions package as a real dependency * refactor: remove more browser globals * test: add browserVersion to make tests more stable + no node:join in specs * chore: prettier * chore: remove old eslint files * fix: baseUrl is mandatory, must fail if missing * fix: load ui5 libs at once * refactor: client-side-js is also typed * fix: use control.getMetadata().getName() * docs: add statement for new bidi protocol * Update docs/migration.md Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> * fix: multiremote race condition where injectUI5 was not done * wip: use stable ubuntu version * wip: decrease ubuntu version * chore: add multiremote to the tests again * refactor: linting * refactor: ts erasableSyntaxOnly * fix: wdio bridge * chore: use stable ubuntu version due to chrome138 pipeline issues * fix: increase timeout for ts tests in pipeline --------- Co-authored-by: Simon Coen <32054457+Siolto@users.noreply.github.com> Co-authored-by: Volker Buzek <dev@metalodge.com> Co-authored-by: Marian Zeis <13335743+marianfoo@users.noreply.github.com> Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> Co-authored-by: Siolto <simon.coen@js-soft.com>

fix: performance

e76d09f

Siolto requested a review from vobu February 11, 2025 23:10

Siolto temporarily deployed to internal February 11, 2025 23:10 — with GitHub Actions Inactive

Siolto mentioned this pull request Feb 12, 2025

feat: wdio v9 compat #645

Closed

Siolto added 10 commits February 12, 2025 18:54

fix: improve error handling and result processing in executeControlMe…

c8c2955

…thod

fix: refactor executeObjectMethod not using callbacks

9537088

test: update DateTimePicker API test to use ISO string format

68a1ec4

fix: update generated-methods test to handle different date formats b…

64b94a0

…ased on protocol

fix: await browser.asControl in press test for proper asynchronous ha…

9b511c4

…ndling

fix: bidi race condition in retrieveElements

594c704

fix: update Planning Calendar test to handle BIDI-specific DOM controls

c61d284

fix: optimize control retrieval for BIDI by removing unnecessary await

a47ea66

fix: adjust BIDI handling in exec test to avoid Promise.all for perfo…

0635b06

…rmance

chore: configuration

c28aaec

Siolto temporarily deployed to internal February 12, 2025 19:53 — with GitHub Actions Inactive

fix: add TODO comment for handling UI5 injection options

657f6d6

Siolto temporarily deployed to internal February 12, 2025 20:12 — with GitHub Actions Inactive

refactor: remove propietary waitForUI5

d8a7dc8

Siolto temporarily deployed to internal April 10, 2025 18:34 — with GitHub Actions Inactive

Siolto marked this pull request as ready for review April 10, 2025 18:35

Siolto requested review from dominikfeininger and marianfoo April 10, 2025 18:37

refactor: run planningCalendar test also with BiDi protocol

60292bf

Siolto temporarily deployed to internal April 17, 2025 08:48 — with GitHub Actions Inactive

fix: improve peformance while retrieving multiple elements (#665)

f0a9c7e

[skip ci]

vobu temporarily deployed to internal May 16, 2025 15:45 — with GitHub Actions Inactive

vobu added 4 commits May 21, 2025 10:03

refactor: rename clientSide__checkForUI5Ready to clientSide_checkForU…

83787c0

…I5Ready for consistency

chore: lock file

8609d21

fix: test

2dcee32

refactor: rm unnecessary await

4d957db

vobu reviewed May 21, 2025

View reviewed changes

vobu temporarily deployed to internal May 21, 2025 12:43 — with GitHub Actions Inactive

vobu approved these changes May 21, 2025

View reviewed changes

Siolto merged commit d63818d into v9-support Jun 4, 2025
1 of 2 checks passed

vobu deleted the v9_performance_fix_0.1 branch July 7, 2025 10:03

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Comments

fix: performance#658

fix: performance#658
Siolto merged 19 commits intov9-supportfrom
v9_performance_fix_0.1

Siolto commented Feb 11, 2025

Uh oh!

Siolto commented Feb 12, 2025 •

edited

Loading

Uh oh!

Siolto commented Feb 12, 2025 •

edited

Loading

Uh oh!

vobu commented Feb 13, 2025 •

edited

Loading

Uh oh!

Siolto commented Feb 24, 2025

Uh oh!

Siolto commented Apr 10, 2025

Uh oh!

vobu commented May 20, 2025

Uh oh!

vobu May 21, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Comments

Conversation

Siolto commented Feb 11, 2025

Uh oh!

Siolto commented Feb 12, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

I call it the Promisified callback hell 🔥 🔥

Uh oh!

Siolto commented Feb 12, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Performance update

old BIDI performance:

new BIDI performance:

old webdriver performance:

new webdriver performance:

Uh oh!

vobu commented Feb 13, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Siolto commented Feb 24, 2025

Uh oh!

Siolto commented Apr 10, 2025

Uh oh!

vobu commented May 20, 2025

Uh oh!

vobu May 21, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Siolto commented Feb 12, 2025 •

edited

Loading

Siolto commented Feb 12, 2025 •

edited

Loading

vobu commented Feb 13, 2025 •

edited

Loading