fix: make localized landing hero resilient to long copy

[masarray]

Goal

Fix the production Indonesian landing-page hero clipping without weakening the English composition or changing content/release behavior.

Root cause

hero-p0.css forced .gradient-text to white-space: nowrap inside a relatively narrow desktop text column. Indonesian Dapat Anda percaya. is longer than the English line and can escape the column; .hero then clips the overflow.

P0 fix

• remove translated headline length from the layout contract: gradient headline may wrap naturally instead of being clipped;
• slightly rebalance the Indonesian desktop hero to 50/50 message/product authority;
• tune Indonesian headline size without making the global English hero smaller;
• allow the long Indonesian eyebrow to wrap safely;
• on compact desktop widths, allow Indonesian CTA/proof rows to wrap deliberately instead of overflowing;
• keep mobile behavior, copy, SEO, download routing, release sync and runtime untouched.

Regression guard

Adds a source regression test that fails if the gradient headline is forced back to white-space: nowrap and verifies the Indonesian desktop safety rules remain present.

Scope

2 files only: landing/hero-p0.css and one regression test.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: a4cd77c53e

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Scope the wrapping assertion to the gradient rule

If .hero h1 .gradient-text is changed back to white-space: nowrap, this regression test still passes: the positive assertion finds white-space: normal in the Indonesian eyebrow rule, while the negative assertion cannot match the current gradient block because max-inline-size appears between display and white-space. Extract or otherwise inspect the gradient rule itself so the advertised clipping regression is actually detected.

Useful? React with 👍 / 👎.