Revert: #33618 Grid Demos csp fix — to verify visual tests regression source

[aleksei-semikozov]

Diagnostic revert of #33618 (Grid Demos: fix csp in treelist-overview and grid-rowselection).

Purpose: confirm that #33618 (merged 2026-05-19 12:51 as commit 685ce5f5) is the root cause of jquery screenshot test regression on 26_1.

Evidence pointing at #33618:

• Before 685ce5f5: all jquery screenshot tests pass stably on master 26_1
• After 685ce5f5: jquery(2/3)-material starts failing — first crack
• Subsequent commits add more flakies, eventually all 4 tests stable-fail after chore(deps): update dependency webpack-dev-server to v5.2.4 [security] #33628 webpack bump (5-21)

Bisection summary (visual-tests-demos.yml on 26_1):

Date	Commit	Title	jquery fails
5-18 ...	up to 36c73f41	various	all green
5-19 10:17	312163ac	CI publish tag	all green
5-19 12:51	685ce5f5	#33618 Grid Demos csp	first ❌ (2/3-material)
5-19 13:56	74714d8e	#33622 AI utils	+1 fail
...	...	...	escalating
5-21 21:11	d9629585	#33628 webpack	stable 4/6 fail

Diagnostic PRs already done:

• Security: update @angular/platform-server, ws, brace-expansion #33669 (our security PR) — 4 jquery fails, baseline
• [test] Empty commit from master to isolate jquery regression #33671 (empty commit from master, closed) — same 4 fails ✅ proves master regression
• Revert: webpack-dev-server 5.2.4 (#33628) — to verify jquery screenshot regression #33675 (revert chore(deps): update dependency webpack-dev-server to v5.2.4 [security] #33628, closed) — STILL 5 jquery fails ❌ proves chore(deps): update dependency webpack-dev-server to v5.2.4 [security] #33628 is not root cause

Expected outcome: if jquery tests pass on this revert PR, #33618 is the root cause and team needs to:

• Re-do Grid Demos: fix csp in treelist-overview and grid-rowselection #33618 differently (without breaking visual tests)
• Or regenerate screenshot baselines properly
• Or accept that we need to revert and re-apply CSP fixes carefully

[aleksei-semikozov]

Diagnostic complete. Reverting #33618 does NOT fix the regression — same 4 jquery tests still fail.

Root cause identified separately: timing issue with remote-data demos (DataGrid/AIColumns hitting OpenAI proxy, DataGrid/RemoteVirtualScrolling + RemoteGrouping + PivotGrid/WebAPIService hitting js.devexpress.com Sales API). Screenshot capture happens before async data finishes loading.

See diff artifacts: 'Loading...' text visible in actual screenshots; affected demos all fetch from external sources.

Likely catalyst: #33639 (Demos: DataGrid/TreeList aicolumn — handle inline module) which changed JS loading timing for AI demos. Real fix needed: increase screenshot wait timeouts, mock external data in CI, or skip AI/Remote demos in visual tests.