Allow t() to return falsy result instead of translation key or provide a way to add a fallback value

[wscourge]

Is your feature request related to a problem? Please describe.

For Next.js 13 Metadata, I want to do the following:

export async function generateMetadata({ params: { locale }) {
  const t = await getTranslator(locale, "home")

  return {
    title: t("title"),
    openGraph: {
      title: t("seo.og.title") || t("seo.title") || t("title"),
    }
  }
}

Describe the solution you'd like

Either will do:

1. t.optional('home') that'd to return "" or other fals-y value. Similar to t.rich().
2. t('seo.og.title', null, t('seo.title', null, t('title'))), where the 3rd arguments is a fallback value.

Describe alternatives you've considered

As of now, I'm doing it with my customized version of the t() function:

export const t = async (locale: Locale, at: string, paths: Array<string>) => {
  const t = await getTranslator(locale, at)
  return t(paths.find((path: string) => !t(path).startsWith(`${at}.`)))
}

// usage
t(locale, "home", ["seo.og.title", "seo.title", "title"])

It works, but results in the pretty annoying, irrelevant error:

IntlError: MISSING_MESSAGE: Could not resolve seo.og.title in site.home.
at getFallbackFromErrorAndNotify (webpack-internal:///(rsc)/./node_modules/use-intl/dist/development/createIntl-dfba4462.js:131:23)
at translateBaseFn (webpack-internal:///(rsc)/./node_modules/use-intl/dist/development/createIntl-dfba4462.js:153:20)
at translateFn (webpack-internal:///(rsc)/./node_modules/use-intl/dist/development/createIntl-dfba4462.js:222:24)
at base (webpack-internal:///(rsc)/./node_modules/use-intl/dist/development/core.js:24:16)
at eval (webpack-internal:///(rsc)/./lib/i18n.ts:26:34)
at Array.find ()
at t (webpack-internal:///(rsc)/./lib/i18n.ts:26:20)
at async Promise.all (index 4)
at async localizedPageMetadata (webpack-internal:///(rsc)/./lib/seo.ts:44:82) {
code: 'MISSING_MESSAGE',
originalMessage: 'Could not resolve seo.tw.title in site.home.'
}

[amannn]

Thanks for the request, this seems to be a duplicate of #88. I'll close this ticket in favor of the existing one.

[wscourge]

Sure, sorry for the dupe

[wscourge]

Is there currently a way to suppress the error tho?

[amannn]

Yep, please see the alternatives section in the linked issue.