@ascribed/astro 0.2.1-next.16 → 0.2.1-next.17

package/dist/project.js.map

@@ -1 +1 @@
-{"version":3,"file":"project.js","sourceRoot":"","sources":["../src/project.ts"],"names":[],"mappings":"AAAA,8EAA8E;AAC9E,uEAAuE;AACvE,oDAAoD;AACpD,yBAAyB;AACzB,OAAO,EAAE,YAAY,EAAE,MAAM,SAAS,CAAC;AACvC,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,EAAE,KAAK,EAAE,MAAM,WAAW,CAAC;AA4BlC,kFAAkF;AAClF,MAAM,UAAU,WAAW,CAAC,GAAW;IACrC,MAAM,UAAU,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE,cAAc,CAAC,CAAC;IAClD,IAAI,KAA8B,CAAC;IACnC,IAAI,CAAC;QACH,KAAK,GAAG,KAAK,CAAC,YAAY,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC,CAAC;IAClD,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,MAAM,IAAI,KAAK,CACb,8BAA8B,UAAU,KAAK,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,EACrG,EAAE,KAAK,EAAE,KAAK,EAAE,CACjB,CAAC;IACJ,CAAC;IACD,MAAM,OAAO,GAAG,OAAO,CAAC,KAAK,EAAE,SAAS,CAAC,CAAC;IAC1C,MAAM,QAAQ,GAAG,OAAO,CAAC,KAAK,EAAE,UAAU,CAAC,CAAC;IAC5C,MAAM,SAAS,GACb,OAAO,OAAO,CAAC,YAAY,CAAC,KAAK,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC,gBAAgB,CAAC;IACvF,MAAM,UAAU,GAAG,OAAO,OAAO,CAAC,cAAc,CAAC,KAAK,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC,cAAc,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IAClG,MAAM,aAAa,GAAG,QAAQ,CAAC,gBAAgB,CAAC,KAAK,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,QAAQ,CAAC;IAClF,OAAO;QACL,GAAG;QACH,UAAU;QACV,WAAW,EAAE,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,UAAU,CAAC;QAC1C,UAAU,EAAE,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,SAAS,CAAC;QACxC,QAAQ,EAAE;YACR,IAAI,EAAE,OAAO,QAAQ,CAAC,MAAM,CAAC,KAAK,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,SAAS;YACzE,QAAQ,EAAE,aAAa,CACrB,OAAO,QAAQ,CAAC,WAAW,CAAC,KAAK,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,GAAG,CACxE;YACD,aAAa;YACb,MAAM,EAAE,QAAQ,CAAC,QAAQ,CAAC,KAAK,IAAI;SACpC;QACD,MAAM,EAAE,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,QAAQ,CAAC,CAAC;QAC7C,QAAQ,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,SAAS,EAAE,KAAK,EAAE,MAAM,CAAC;QAChE,SAAS,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,SAAS,EAAE,KAAK,EAAE,OAAO,CAAC;KACnE,CAAC;AACJ,CAAC;AAED,SAAS,OAAO,CAAC,KAA8B,EAAE,IAAY;IAC3D,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC;IAC1B,OAAO,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,KAAK,IAAI,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC;QACzE,CAAC,CAAE,KAAiC;QACpC,CAAC,CAAC,EAAE,CAAC;AACT,CAAC;AAED,iFAAiF;AACjF,MAAM,UAAU,aAAa,CAAC,IAAY;IACxC,IAAI,MAAM,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IACzB,IAAI,CAAC,MAAM,CAAC,UAAU,CAAC,GAAG,CAAC;QAAE,MAAM,GAAG,IAAI,MAAM,EAAE,CAAC;IACnD,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,GAAG,CAAC;QAAE,MAAM,IAAI,GAAG,CAAC;IACzC,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,SAAS,CAAC,IAAY,EAAE,aAA4C;IAClF,MAAM,IAAI,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;IACjC,OAAO,aAAa,KAAK,OAAO,IAAI,IAAI,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;AAC9E,CAAC;AASD;;;;;;;;;;;GAWG;AACH,MAAM,UAAU,aAAa,CAC3B,QAAgC,EAChC,KAAmB,EACnB,UAAmB;IAEnB,MAAM,QAAQ,GAA0B,EAAE,CAAC;IAC3C,IAAI,KAAK,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,MAAM,IAAI,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,QAAQ,CAAC,IAAI,KAAK,SAAS,CAAC,EAAE,QAAQ,CAAC,IAAI,CAAC;QAC5F,IAAI,IAAI,KAAK,SAAS;YAAE,QAAQ,CAAC,IAAI,GAAG,IAAI,CAAC;IAC/C,CAAC;IACD,MAAM,KAAK,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;IAC1B,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,EAAE,OAAO,EAAE,KAAK,EAAE,QAAQ,EAAE,CAAC;IAC7D,IAAI,KAAK,CAAC,aAAa,KAAK,QAAQ;QAAE,QAAQ,CAAC,aAAa,GAAG,KAAK,CAAC,QAAQ,CAAC,aAAa,CAAC;IAC5F,MAAM,aAAa,GAAG,QAAQ,CAAC,aAAa,IAAI,KAAK,CAAC,aAAa,CAAC;IACpE,MAAM,IAAI,GAAG,SAAS,CACpB,UAAU,IAAI,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,KAAK,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,EACtF,aAAa,CACd,CAAC;IACF,IAAI,IAAI,KAAK,KAAK,CAAC,IAAI;QAAE,QAAQ,CAAC,IAAI,GAAG,IAAI,CAAC;IAC9C,OAAO,EAAE,OAAO,EAAE,EAAE,GAAG,KAAK,EAAE,GAAG,QAAQ,EAAE,EAAE,QAAQ,EAAE,CAAC;AAC1D,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,kBAAkB,CAChC,OAAoB,EACpB,KAAwE;IAExE,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,MAAM,EAAE,QAAQ,EAAE,aAAa,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC,QAAQ,CAAC;IAC3D,oGAAoG;IACpG,IAAI,CAAC,QAAQ,CAAC,UAAU,CAAC,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC;QACpD,QAAQ,CAAC,IAAI,CACX,4BAA4B,QAAQ,+BAA+B,KAAK,CAAC,IAAI,mDAAmD,CACjI,CAAC;IACJ,CAAC;IACD,0EAA0E;IAC1E,IAAI,KAAK,CAAC,aAAa,KAAK,QAAQ,IAAI,KAAK,CAAC,aAAa,KAAK,aAAa,EAAE,CAAC;QAC9E,QAAQ,CAAC,IAAI,CACX,iCAAiC,aAAa,wCAAwC,KAAK,CAAC,aAAa,IAAI,CAC9G,CAAC;IACJ,CAAC;IACD,IAAI,IAAI,KAAK,SAAS,IAAI,KAAK,CAAC,IAAI,KAAK,SAAS,IAAI,MAAM,CAAC,IAAI,CAAC,KAAK,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC;QAC1F,QAAQ,CAAC,IAAI,CAAC,uBAAuB,IAAI,+BAA+B,KAAK,CAAC,IAAI,IAAI,CAAC,CAAC;IAC1F,CAAC;IACD,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,SAAS,MAAM,CAAC,GAAW;IACzB,IAAI,CAAC;QACH,OAAO,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC;IAC7B,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,GAAG,CAAC;IACb,CAAC;AACH,CAAC","sourcesContent":["// What the integration needs from `ascribe.toml`: where a build's site output\n// goes, and the consumer settings that must agree with `astro.config`.\n// The compiler validates the whole file; this reads\n// only what it compiled.\nimport { readFileSync } from \"node:fs\";\nimport path from \"node:path\";\nimport { parse } from \"smol-toml\";\n\n/** The parts of an Ascribe project the integration reads. */\nexport interface ProjectInfo {\n  /** The directory holding `ascribe.toml`. */\n  dir: string;\n  /** The path of `ascribe.toml`. */\n  configPath: string;\n  /** Source files and assets watched during development. */\n  contentRoot: string;\n  /** Generated output, never treated as a source change. */\n  outputRoot: string;\n  /** `[consumer]`, with the profile's defaults. */\n  consumer: {\n    site: string | undefined;\n    basePath: string;\n    trailingSlash: \"always\" | \"never\";\n    /** Whether builds publish what agents read: `llms.txt` and each page's Markdown, in the plain output. */\n    agents: boolean;\n  };\n  /** The names of `[builds]`'s tables, in order: empty when there are none, and the one build is `site`. */\n  builds: string[];\n  /** The site output's root for a build: `<output-dir>/<build>/site`. */\n  siteRoot(build: string): string;\n  /** The plain output's root for a build: `<output-dir>/<build>/plain`. */\n  plainRoot(build: string): string;\n}\n\n/** Reads `ascribe.toml` in `dir`. Throws a readable error if it can't be read. */\nexport function readProject(dir: string): ProjectInfo {\n  const configPath = path.join(dir, \"ascribe.toml\");\n  let table: Record<string, unknown>;\n  try {\n    table = parse(readFileSync(configPath, \"utf8\"));\n  } catch (error) {\n    throw new Error(\n      `@ascribed/astro can't read ${configPath}: ${error instanceof Error ? error.message : String(error)}`,\n      { cause: error },\n    );\n  }\n  const project = section(table, \"project\");\n  const consumer = section(table, \"consumer\");\n  const outputDir =\n    typeof project[\"output-dir\"] === \"string\" ? project[\"output-dir\"] : \".ascribe/build\";\n  const contentDir = typeof project[\"content-root\"] === \"string\" ? project[\"content-root\"] : \"docs\";\n  const trailingSlash = consumer[\"trailing-slash\"] === \"never\" ? \"never\" : \"always\";\n  return {\n    dir,\n    configPath,\n    contentRoot: path.resolve(dir, contentDir),\n    outputRoot: path.resolve(dir, outputDir),\n    consumer: {\n      site: typeof consumer[\"site\"] === \"string\" ? consumer[\"site\"] : undefined,\n      basePath: normalizeBase(\n        typeof consumer[\"base-path\"] === \"string\" ? consumer[\"base-path\"] : \"/\",\n      ),\n      trailingSlash,\n      agents: consumer[\"agents\"] === true,\n    },\n    builds: Object.keys(section(table, \"builds\")),\n    siteRoot: (build) => path.resolve(dir, outputDir, build, \"site\"),\n    plainRoot: (build) => path.resolve(dir, outputDir, build, \"plain\"),\n  };\n}\n\nfunction section(table: Record<string, unknown>, name: string): Record<string, unknown> {\n  const value = table[name];\n  return typeof value === \"object\" && value !== null && !Array.isArray(value)\n    ? (value as Record<string, unknown>)\n    : {};\n}\n\n/** A base path with a leading and a trailing `/`, as Ascribe's router has it. */\nexport function normalizeBase(base: string): string {\n  let result = base.trim();\n  if (!result.startsWith(\"/\")) result = `/${result}`;\n  if (!result.endsWith(\"/\")) result += \"/\";\n  return result;\n}\n\n/**\n * The route of a base path, ending the way `trailingSlash` says the site's\n * routes do: the site's root for Astro's `base`, a project's root for its\n * `[consumer] base-path`. (`import.meta.env.BASE_URL` isn't replaced in a\n * package's components.)\n */\nexport function homeRoute(base: string, trailingSlash: \"always\" | \"never\" | \"ignore\"): string {\n  const root = normalizeBase(base);\n  return trailingSlash === \"never\" && root !== \"/\" ? root.slice(0, -1) : root;\n}\n\n/** Astro's routing settings, as the integration reads and supplies them. */\nexport interface AstroRouting {\n  base: string;\n  trailingSlash: \"always\" | \"never\" | \"ignore\";\n  site: string | undefined;\n}\n\n/**\n * Astro's routing with what astro.config leaves unset taken from `[consumer]`,\n * and the settings taken, for `updateConfig`. Astro gives an integration its\n * configuration with its defaults filled in, so unset is the default: no\n * `site`, `trailingSlash: \"ignore\"`, and `base: \"/\"`. `site` and\n * `trailingSlash` come from the first project that has them, since every\n * project must agree with them anyway. `base` comes from the first project's\n * `base-path` only with `supplyBase`, as with one project; with several, it\n * stays the site's, the root every base path is under. A base Astro has is\n * rewritten to end the way a supplied `trailingSlash` says, as Astro writes\n * it when the setting is its own.\n */\nexport function supplyRouting(\n  projects: readonly ProjectInfo[],\n  astro: AstroRouting,\n  supplyBase: boolean,\n): { routing: AstroRouting; supplied: Partial<AstroRouting> } {\n  const supplied: Partial<AstroRouting> = {};\n  if (astro.site === undefined) {\n    const site = projects.find((project) => project.consumer.site !== undefined)?.consumer.site;\n    if (site !== undefined) supplied.site = site;\n  }\n  const first = projects[0];\n  if (first === undefined) return { routing: astro, supplied };\n  if (astro.trailingSlash === \"ignore\") supplied.trailingSlash = first.consumer.trailingSlash;\n  const trailingSlash = supplied.trailingSlash ?? astro.trailingSlash;\n  const base = homeRoute(\n    supplyBase && normalizeBase(astro.base) === \"/\" ? first.consumer.basePath : astro.base,\n    trailingSlash,\n  );\n  if (base !== astro.base) supplied.base = base;\n  return { routing: { ...astro, ...supplied }, supplied };\n}\n\n/**\n * The ways `ascribe.toml`'s `[consumer]` disagrees with Astro's configuration,\n * as sentences. Ascribe writes every link with its own settings, so a site\n * whose routes differ would have broken links.\n */\nexport function consumerMismatches(\n  project: ProjectInfo,\n  astro: { base: string; trailingSlash: string; site: string | undefined },\n): string[] {\n  const problems: string[] = [];\n  const { basePath, trailingSlash, site } = project.consumer;\n  // A base path under Astro's `base` leaves the rest of the site to other pages, such as a home page.\n  if (!basePath.startsWith(normalizeBase(astro.base))) {\n    problems.push(\n      `[consumer] base-path is \"${basePath}\", but Astro's \\`base\\` is \"${astro.base}\". Set base-path to that path or a path under it.`,\n    );\n  }\n  // Astro's \"ignore\" serves both forms, so Ascribe's links work either way.\n  if (astro.trailingSlash !== \"ignore\" && astro.trailingSlash !== trailingSlash) {\n    problems.push(\n      `[consumer] trailing-slash is \"${trailingSlash}\", but Astro's \\`trailingSlash\\` is \"${astro.trailingSlash}\".`,\n    );\n  }\n  if (site !== undefined && astro.site !== undefined && origin(site) !== origin(astro.site)) {\n    problems.push(`[consumer] site is \"${site}\", but Astro's \\`site\\` is \"${astro.site}\".`);\n  }\n  return problems;\n}\n\nfunction origin(url: string): string {\n  try {\n    return new URL(url).origin;\n  } catch {\n    return url;\n  }\n}\n"]}
+{"version":3,"file":"project.js","sourceRoot":"","sources":["../src/project.ts"],"names":[],"mappings":"AAAA,8EAA8E;AAC9E,uEAAuE;AACvE,oDAAoD;AACpD,yBAAyB;AACzB,OAAO,EAAE,YAAY,EAAE,MAAM,SAAS,CAAC;AACvC,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,EAAE,KAAK,EAAE,MAAM,WAAW,CAAC;AA8BlC,kFAAkF;AAClF,MAAM,UAAU,WAAW,CAAC,GAAW;IACrC,MAAM,UAAU,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE,cAAc,CAAC,CAAC;IAClD,IAAI,KAA8B,CAAC;IACnC,IAAI,CAAC;QACH,KAAK,GAAG,KAAK,CAAC,YAAY,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC,CAAC;IAClD,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,MAAM,IAAI,KAAK,CACb,8BAA8B,UAAU,KAAK,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,EACrG,EAAE,KAAK,EAAE,KAAK,EAAE,CACjB,CAAC;IACJ,CAAC;IACD,MAAM,OAAO,GAAG,OAAO,CAAC,KAAK,EAAE,SAAS,CAAC,CAAC;IAC1C,MAAM,QAAQ,GAAG,OAAO,CAAC,KAAK,EAAE,UAAU,CAAC,CAAC;IAC5C,MAAM,SAAS,GACb,OAAO,OAAO,CAAC,YAAY,CAAC,KAAK,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC,gBAAgB,CAAC;IACvF,MAAM,UAAU,GAAG,OAAO,OAAO,CAAC,cAAc,CAAC,KAAK,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC,cAAc,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IAClG,MAAM,aAAa,GAAG,QAAQ,CAAC,gBAAgB,CAAC,KAAK,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,QAAQ,CAAC;IAClF,OAAO;QACL,GAAG;QACH,UAAU;QACV,WAAW,EAAE,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,UAAU,CAAC;QAC1C,UAAU,EAAE,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,SAAS,CAAC;QACxC,QAAQ,EAAE;YACR,IAAI,EAAE,OAAO,QAAQ,CAAC,MAAM,CAAC,KAAK,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,SAAS;YACzE,QAAQ,EAAE,aAAa,CACrB,OAAO,QAAQ,CAAC,WAAW,CAAC,KAAK,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,GAAG,CACxE;YACD,aAAa;YACb,MAAM,EAAE,QAAQ,CAAC,QAAQ,CAAC,KAAK,IAAI;YACnC,OAAO,EAAE,QAAQ,CAAC,SAAS,CAAC,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,SAAS;SACnE;QACD,MAAM,EAAE,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,QAAQ,CAAC,CAAC;QAC7C,QAAQ,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,SAAS,EAAE,KAAK,EAAE,MAAM,CAAC;QAChE,SAAS,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,SAAS,EAAE,KAAK,EAAE,OAAO,CAAC;KACnE,CAAC;AACJ,CAAC;AAED,SAAS,OAAO,CAAC,KAA8B,EAAE,IAAY;IAC3D,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC;IAC1B,OAAO,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,KAAK,IAAI,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC;QACzE,CAAC,CAAE,KAAiC;QACpC,CAAC,CAAC,EAAE,CAAC;AACT,CAAC;AAED,iFAAiF;AACjF,MAAM,UAAU,aAAa,CAAC,IAAY;IACxC,IAAI,MAAM,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IACzB,IAAI,CAAC,MAAM,CAAC,UAAU,CAAC,GAAG,CAAC;QAAE,MAAM,GAAG,IAAI,MAAM,EAAE,CAAC;IACnD,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,GAAG,CAAC;QAAE,MAAM,IAAI,GAAG,CAAC;IACzC,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,SAAS,CAAC,IAAY,EAAE,aAA4C;IAClF,MAAM,IAAI,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;IACjC,OAAO,aAAa,KAAK,OAAO,IAAI,IAAI,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;AAC9E,CAAC;AASD;;;;;;;;;;;GAWG;AACH,MAAM,UAAU,aAAa,CAC3B,QAAgC,EAChC,KAAmB,EACnB,UAAmB;IAEnB,MAAM,QAAQ,GAA0B,EAAE,CAAC;IAC3C,IAAI,KAAK,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,MAAM,IAAI,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,QAAQ,CAAC,IAAI,KAAK,SAAS,CAAC,EAAE,QAAQ,CAAC,IAAI,CAAC;QAC5F,IAAI,IAAI,KAAK,SAAS;YAAE,QAAQ,CAAC,IAAI,GAAG,IAAI,CAAC;IAC/C,CAAC;IACD,MAAM,KAAK,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;IAC1B,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,EAAE,OAAO,EAAE,KAAK,EAAE,QAAQ,EAAE,CAAC;IAC7D,IAAI,KAAK,CAAC,aAAa,KAAK,QAAQ;QAAE,QAAQ,CAAC,aAAa,GAAG,KAAK,CAAC,QAAQ,CAAC,aAAa,CAAC;IAC5F,MAAM,aAAa,GAAG,QAAQ,CAAC,aAAa,IAAI,KAAK,CAAC,aAAa,CAAC;IACpE,MAAM,IAAI,GAAG,SAAS,CACpB,UAAU,IAAI,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,KAAK,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,EACtF,aAAa,CACd,CAAC;IACF,IAAI,IAAI,KAAK,KAAK,CAAC,IAAI;QAAE,QAAQ,CAAC,IAAI,GAAG,IAAI,CAAC;IAC9C,OAAO,EAAE,OAAO,EAAE,EAAE,GAAG,KAAK,EAAE,GAAG,QAAQ,EAAE,EAAE,QAAQ,EAAE,CAAC;AAC1D,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,kBAAkB,CAChC,OAAoB,EACpB,KAAwE;IAExE,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,MAAM,EAAE,QAAQ,EAAE,aAAa,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC,QAAQ,CAAC;IAC3D,oGAAoG;IACpG,IAAI,CAAC,QAAQ,CAAC,UAAU,CAAC,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC;QACpD,QAAQ,CAAC,IAAI,CACX,4BAA4B,QAAQ,+BAA+B,KAAK,CAAC,IAAI,mDAAmD,CACjI,CAAC;IACJ,CAAC;IACD,0EAA0E;IAC1E,IAAI,KAAK,CAAC,aAAa,KAAK,QAAQ,IAAI,KAAK,CAAC,aAAa,KAAK,aAAa,EAAE,CAAC;QAC9E,QAAQ,CAAC,IAAI,CACX,iCAAiC,aAAa,wCAAwC,KAAK,CAAC,aAAa,IAAI,CAC9G,CAAC;IACJ,CAAC;IACD,IAAI,IAAI,KAAK,SAAS,IAAI,KAAK,CAAC,IAAI,KAAK,SAAS,IAAI,MAAM,CAAC,IAAI,CAAC,KAAK,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC;QAC1F,QAAQ,CAAC,IAAI,CAAC,uBAAuB,IAAI,+BAA+B,KAAK,CAAC,IAAI,IAAI,CAAC,CAAC;IAC1F,CAAC;IACD,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,SAAS,MAAM,CAAC,GAAW;IACzB,IAAI,CAAC;QACH,OAAO,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC;IAC7B,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,GAAG,CAAC;IACb,CAAC;AACH,CAAC","sourcesContent":["// What the integration needs from `ascribe.toml`: where a build's site output\n// goes, and the consumer settings that must agree with `astro.config`.\n// The compiler validates the whole file; this reads\n// only what it compiled.\nimport { readFileSync } from \"node:fs\";\nimport path from \"node:path\";\nimport { parse } from \"smol-toml\";\n\n/** The parts of an Ascribe project the integration reads. */\nexport interface ProjectInfo {\n  /** The directory holding `ascribe.toml`. */\n  dir: string;\n  /** The path of `ascribe.toml`. */\n  configPath: string;\n  /** Source files and assets watched during development. */\n  contentRoot: string;\n  /** Generated output, never treated as a source change. */\n  outputRoot: string;\n  /** `[consumer]`, with the profile's defaults. */\n  consumer: {\n    site: string | undefined;\n    basePath: string;\n    trailingSlash: \"always\" | \"never\";\n    /** Whether builds publish what agents read: `llms.txt` and each page's Markdown, in the plain output. */\n    agents: boolean;\n    /** The host whose format the site's response headers are written in: `\"netlify\"`, or none. */\n    headers: \"netlify\" | undefined;\n  };\n  /** The names of `[builds]`'s tables, in order: empty when there are none, and the one build is `site`. */\n  builds: string[];\n  /** The site output's root for a build: `<output-dir>/<build>/site`. */\n  siteRoot(build: string): string;\n  /** The plain output's root for a build: `<output-dir>/<build>/plain`. */\n  plainRoot(build: string): string;\n}\n\n/** Reads `ascribe.toml` in `dir`. Throws a readable error if it can't be read. */\nexport function readProject(dir: string): ProjectInfo {\n  const configPath = path.join(dir, \"ascribe.toml\");\n  let table: Record<string, unknown>;\n  try {\n    table = parse(readFileSync(configPath, \"utf8\"));\n  } catch (error) {\n    throw new Error(\n      `@ascribed/astro can't read ${configPath}: ${error instanceof Error ? error.message : String(error)}`,\n      { cause: error },\n    );\n  }\n  const project = section(table, \"project\");\n  const consumer = section(table, \"consumer\");\n  const outputDir =\n    typeof project[\"output-dir\"] === \"string\" ? project[\"output-dir\"] : \".ascribe/build\";\n  const contentDir = typeof project[\"content-root\"] === \"string\" ? project[\"content-root\"] : \"docs\";\n  const trailingSlash = consumer[\"trailing-slash\"] === \"never\" ? \"never\" : \"always\";\n  return {\n    dir,\n    configPath,\n    contentRoot: path.resolve(dir, contentDir),\n    outputRoot: path.resolve(dir, outputDir),\n    consumer: {\n      site: typeof consumer[\"site\"] === \"string\" ? consumer[\"site\"] : undefined,\n      basePath: normalizeBase(\n        typeof consumer[\"base-path\"] === \"string\" ? consumer[\"base-path\"] : \"/\",\n      ),\n      trailingSlash,\n      agents: consumer[\"agents\"] === true,\n      headers: consumer[\"headers\"] === \"netlify\" ? \"netlify\" : undefined,\n    },\n    builds: Object.keys(section(table, \"builds\")),\n    siteRoot: (build) => path.resolve(dir, outputDir, build, \"site\"),\n    plainRoot: (build) => path.resolve(dir, outputDir, build, \"plain\"),\n  };\n}\n\nfunction section(table: Record<string, unknown>, name: string): Record<string, unknown> {\n  const value = table[name];\n  return typeof value === \"object\" && value !== null && !Array.isArray(value)\n    ? (value as Record<string, unknown>)\n    : {};\n}\n\n/** A base path with a leading and a trailing `/`, as Ascribe's router has it. */\nexport function normalizeBase(base: string): string {\n  let result = base.trim();\n  if (!result.startsWith(\"/\")) result = `/${result}`;\n  if (!result.endsWith(\"/\")) result += \"/\";\n  return result;\n}\n\n/**\n * The route of a base path, ending the way `trailingSlash` says the site's\n * routes do: the site's root for Astro's `base`, a project's root for its\n * `[consumer] base-path`. (`import.meta.env.BASE_URL` isn't replaced in a\n * package's components.)\n */\nexport function homeRoute(base: string, trailingSlash: \"always\" | \"never\" | \"ignore\"): string {\n  const root = normalizeBase(base);\n  return trailingSlash === \"never\" && root !== \"/\" ? root.slice(0, -1) : root;\n}\n\n/** Astro's routing settings, as the integration reads and supplies them. */\nexport interface AstroRouting {\n  base: string;\n  trailingSlash: \"always\" | \"never\" | \"ignore\";\n  site: string | undefined;\n}\n\n/**\n * Astro's routing with what astro.config leaves unset taken from `[consumer]`,\n * and the settings taken, for `updateConfig`. Astro gives an integration its\n * configuration with its defaults filled in, so unset is the default: no\n * `site`, `trailingSlash: \"ignore\"`, and `base: \"/\"`. `site` and\n * `trailingSlash` come from the first project that has them, since every\n * project must agree with them anyway. `base` comes from the first project's\n * `base-path` only with `supplyBase`, as with one project; with several, it\n * stays the site's, the root every base path is under. A base Astro has is\n * rewritten to end the way a supplied `trailingSlash` says, as Astro writes\n * it when the setting is its own.\n */\nexport function supplyRouting(\n  projects: readonly ProjectInfo[],\n  astro: AstroRouting,\n  supplyBase: boolean,\n): { routing: AstroRouting; supplied: Partial<AstroRouting> } {\n  const supplied: Partial<AstroRouting> = {};\n  if (astro.site === undefined) {\n    const site = projects.find((project) => project.consumer.site !== undefined)?.consumer.site;\n    if (site !== undefined) supplied.site = site;\n  }\n  const first = projects[0];\n  if (first === undefined) return { routing: astro, supplied };\n  if (astro.trailingSlash === \"ignore\") supplied.trailingSlash = first.consumer.trailingSlash;\n  const trailingSlash = supplied.trailingSlash ?? astro.trailingSlash;\n  const base = homeRoute(\n    supplyBase && normalizeBase(astro.base) === \"/\" ? first.consumer.basePath : astro.base,\n    trailingSlash,\n  );\n  if (base !== astro.base) supplied.base = base;\n  return { routing: { ...astro, ...supplied }, supplied };\n}\n\n/**\n * The ways `ascribe.toml`'s `[consumer]` disagrees with Astro's configuration,\n * as sentences. Ascribe writes every link with its own settings, so a site\n * whose routes differ would have broken links.\n */\nexport function consumerMismatches(\n  project: ProjectInfo,\n  astro: { base: string; trailingSlash: string; site: string | undefined },\n): string[] {\n  const problems: string[] = [];\n  const { basePath, trailingSlash, site } = project.consumer;\n  // A base path under Astro's `base` leaves the rest of the site to other pages, such as a home page.\n  if (!basePath.startsWith(normalizeBase(astro.base))) {\n    problems.push(\n      `[consumer] base-path is \"${basePath}\", but Astro's \\`base\\` is \"${astro.base}\". Set base-path to that path or a path under it.`,\n    );\n  }\n  // Astro's \"ignore\" serves both forms, so Ascribe's links work either way.\n  if (astro.trailingSlash !== \"ignore\" && astro.trailingSlash !== trailingSlash) {\n    problems.push(\n      `[consumer] trailing-slash is \"${trailingSlash}\", but Astro's \\`trailingSlash\\` is \"${astro.trailingSlash}\".`,\n    );\n  }\n  if (site !== undefined && astro.site !== undefined && origin(site) !== origin(astro.site)) {\n    problems.push(`[consumer] site is \"${site}\", but Astro's \\`site\\` is \"${astro.site}\".`);\n  }\n  return problems;\n}\n\nfunction origin(url: string): string {\n  try {\n    return new URL(url).origin;\n  } catch {\n    return url;\n  }\n}\n"]}

package/dist/records.d.ts

@@ -1,5 +1,18 @@
 /** How a page is treated: listed, listed and not to be quoted, or not listed. */
 export type Listing = "listed" | "unquoted" | "unlisted";
+/**
+ * Where a page's listing comes from: `default` when nothing sets it, `page`
+ * for the page's own, and for a rule of `[listing]` its level (`site`,
+ * `folder`, or `type`), its folder or type, the rule as written, and its line
+ * in `ascribe.toml`.
+ */
+export interface ListingFrom {
+    level: "default" | "page" | "site" | "folder" | "type";
+    folder?: string;
+    type?: string;
+    rule?: string;
+    line?: number;
+}
 /** One page's record, with the route it's keyed by. */
 export interface PageRecord {
     /** The page's route, from the host's root. */
@@ -13,17 +26,52 @@ export interface PageRecord {
     /** The page's own address, absolute, when the project sets `[consumer] site`. */
     canonical?: string;
     listing: Listing;
+    /** Where `listing` comes from: the page, a rule of `[listing]`, or neither. */
+    listing_from: ListingFrom;
     /** The address of the page's Markdown version, from the host's root, when the build publishes one. */
     markdown?: string;
+    /**
+     * The page's place in the site, from the top down to the page itself: the
+     * page at each of its route's folders that has one, then the page. Absent
+     * in records an earlier Ascribe wrote.
+     */
+    breadcrumbs?: Step[];
+}
+/** One step of a page's breadcrumbs: a page's route, from the host's root, and its title. */
+export interface Step {
+    route: string;
+    title: string;
+}
+/** What a kind of crawler is for: gathering training data, building an AI search index, or fetching at a person's request. */
+type Purpose = "training" | "search" | "on-request";
+/** The group of `robots.txt` a disallowed purpose writes. */
+export interface RobotsGroup {
+    purpose: Purpose;
+    /** Every token of the purpose the binary knows, each a `User-agent` line. */
+    user_agents: string[];
+    /** The group's rules: `Disallow: /`. */
+    lines: string[];
+}
+/** What a project says to crawlers. */
+export interface Robots {
+    /** The project's own lines, with paths from the host's root. */
+    lines: string[];
+    /** What the project says of each purpose it sets. */
+    purposes?: Partial<Record<Purpose, "allow" | "disallow">>;
+    /** A group for each purpose the project disallows. */
+    groups?: RobotsGroup[];
 }
 /** `_ascribe/pages.json`. */
 export interface PageRecords {
     version: number;
     site?: string;
     pages: Record<string, Omit<PageRecord, "route">>;
-    robots: {
-        lines: string[];
-    };
+    robots: Robots;
+}
+/** What one project says to crawlers, by the name of its first entry. */
+export interface ProjectRobots {
+    name: string;
+    robots: Robots;
 }
 /** The records in a build's site output, or `undefined` when there aren't any. */
 export declare function readRecords(siteRoot: string): PageRecords | undefined;
@@ -68,11 +116,21 @@ export declare function sitemapAddresses(sources: SitemapSources): string[];
 /** A sitemap of `addresses`, with no dates. */
 export declare function sitemapXml(addresses: readonly string[]): string;
 /**
- * A site's `robots.txt` from each project's lines: a `User-agent: *` group
- * with every project's lines before its first `User-agent` line, which are
- * for every crawler, then the groups each project wrote whole, then the
- * sitemap's address. `undefined` when it would say nothing.
+ * A site's `robots.txt` from what each project says to crawlers: a group for
+ * each purpose a project disallows, once for the site, then a
+ * `User-agent: *` group with every project's lines before its first
+ * `User-agent` line, which are for every crawler, then the groups each
+ * project wrote whole, then the sitemap's address. `undefined` when it would
+ * say nothing.
+ */
+export declare function robotsText(projects: readonly Robots[], sitemap: string | undefined): string | undefined;
+/**
+ * Two projects that say contradictory things about one purpose, as
+ * sentences: one allows it and the other disallows it. A site has one
+ * `robots.txt`, so they must agree. A project that says nothing of a purpose
+ * agrees with either.
  */
-export declare function robotsText(projects: readonly (readonly string[])[], sitemap: string | undefined): string | undefined;
+export declare function purposeConflicts(projects: readonly ProjectRobots[]): string[];
 /** Where the sitemap is, from the host's root: at Astro's `base`. */
 export declare function sitemapRoute(base: string): string;
+export {};

package/dist/records.js

@@ -122,21 +122,33 @@ export function sitemapXml(addresses) {
 }
 const USER_AGENT = /^\s*user-agent\s*:/i;
 /**
- * A site's `robots.txt` from each project's lines: a `User-agent: *` group
- * with every project's lines before its first `User-agent` line, which are
- * for every crawler, then the groups each project wrote whole, then the
- * sitemap's address. `undefined` when it would say nothing.
+ * A site's `robots.txt` from what each project says to crawlers: a group for
+ * each purpose a project disallows, once for the site, then a
+ * `User-agent: *` group with every project's lines before its first
+ * `User-agent` line, which are for every crawler, then the groups each
+ * project wrote whole, then the sitemap's address. `undefined` when it would
+ * say nothing.
  */
 export function robotsText(projects, sitemap) {
+    const purposes = new Map();
     const everyone = [];
     const groups = [];
-    for (const lines of projects) {
+    for (const robots of projects) {
+        for (const group of robots.groups ?? []) {
+            if (purposes.has(group.purpose))
+                continue;
+            purposes.set(group.purpose, [
+                ...group.user_agents.map((agent) => `User-agent: ${agent}`),
+                ...group.lines,
+            ]);
+        }
+        const lines = robots.lines;
         const first = lines.findIndex((line) => USER_AGENT.test(line));
         everyone.push(...(first === -1 ? lines : lines.slice(0, first)));
         if (first !== -1)
             groups.push(lines.slice(first));
     }
-    const blocks = [];
+    const blocks = [...purposes.values()];
     if (everyone.length > 0)
         blocks.push(["User-agent: *", ...everyone]);
     blocks.push(...groups);
@@ -146,6 +158,23 @@ export function robotsText(projects, sitemap) {
         return undefined;
     return `${blocks.map((block) => block.join("\n")).join("\n\n")}\n`;
 }
+/**
+ * Two projects that say contradictory things about one purpose, as
+ * sentences: one allows it and the other disallows it. A site has one
+ * `robots.txt`, so they must agree. A project that says nothing of a purpose
+ * agrees with either.
+ */
+export function purposeConflicts(projects) {
+    const problems = [];
+    for (const purpose of ["training", "search", "on-request"]) {
+        const allow = projects.find((p) => p.robots.purposes?.[purpose] === "allow");
+        const disallow = projects.find((p) => p.robots.purposes?.[purpose] === "disallow");
+        if (allow === undefined || disallow === undefined)
+            continue;
+        problems.push(`"${disallow.name}" has [crawlers] ${purpose} = "disallow" and "${allow.name}" has ${purpose} = "allow". The site has one robots.txt, so its projects must agree: make them the same, or remove one.`);
+    }
+    return problems;
+}
 /** Where the sitemap is, from the host's root: at Astro's `base`. */
 export function sitemapRoute(base) {
     return `${normalizeBase(base)}sitemap.xml`;

package/dist/records.js.map

@@ -1 +1 @@
-{"version":3,"file":"records.js","sourceRoot":"","sources":["../src/records.ts"],"names":[],"mappings":"AAAA,mEAAmE;AACnE,8EAA8E;AAC9E,2EAA2E;AAC3E,yEAAyE;AACzE,wCAAwC;AACxC,OAAO,EAAE,YAAY,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAC;AACjD,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,EAAE,aAAa,EAAE,MAAM,cAAc,CAAC;AAE7C,8CAA8C;AAC9C,MAAM,OAAO,GAAG,CAAC,CAAC;AA8BlB,SAAS,WAAW,CAAC,QAAgB;IACnC,OAAO,IAAI,CAAC,IAAI,CAAC,QAAQ,EAAE,UAAU,EAAE,YAAY,CAAC,CAAC;AACvD,CAAC;AAED,kFAAkF;AAClF,MAAM,UAAU,WAAW,CAAC,QAAgB;IAC1C,MAAM,IAAI,GAAG,WAAW,CAAC,QAAQ,CAAC,CAAC;IACnC,IAAI,IAAY,CAAC;IACjB,IAAI,CAAC;QACH,IAAI,GAAG,YAAY,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;IACpC,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,MAAM,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAgB,CAAC;IAChD,IAAI,OAAO,CAAC,OAAO,KAAK,OAAO,EAAE,CAAC;QAChC,MAAM,IAAI,KAAK,CACb,oBAAoB,IAAI,eAAe,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,2EAA2E,OAAO,sDAAsD,CACvM,CAAC;IACJ,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,qFAAqF;AACrF,MAAM,QAAQ,GAAG,IAAI,GAAG,EAA+D,CAAC;AAExF;;;;GAIG;AACH,MAAM,UAAU,SAAS,CAAC,QAAgB,EAAE,MAAc;IACxD,IAAI,OAAe,CAAC;IACpB,IAAI,CAAC;QACH,OAAO,GAAG,QAAQ,CAAC,WAAW,CAAC,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC;IACpD,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,IAAI,KAAK,GAAG,QAAQ,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;IACnC,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,CAAC,OAAO,KAAK,OAAO,EAAE,CAAC;QACrD,MAAM,KAAK,GAAG,IAAI,GAAG,EAAsB,CAAC;QAC5C,KAAK,MAAM,MAAM,IAAI,QAAQ,CAAC,WAAW,CAAC,QAAQ,CAAC,CAAC,EAAE,CAAC;YACrD,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,MAAM,CAAC;gBAAE,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;QAClE,CAAC;QACD,KAAK,GAAG,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC;QAC3B,QAAQ,CAAC,GAAG,CAAC,QAAQ,EAAE,KAAK,CAAC,CAAC;IAChC,CAAC;IACD,OAAO,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;AACjC,CAAC;AAED,oEAAoE;AACpE,MAAM,UAAU,QAAQ,CAAC,OAAgC;IACvD,OAAO,MAAM,CAAC,OAAO,CAAC,OAAO,EAAE,KAAK,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,KAAK,EAAE,IAAI,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,GAAG,IAAI,EAAE,CAAC,CAAC,CAAC;AAC3F,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,QAAQ,CAAC,KAAa;IACpC,MAAM,SAAS,GAAG,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,EAAE,CAAC;IAC9D,OAAO,SAAS,CAAC,MAAM,GAAG,CAAC,IAAI,SAAS,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;AAC9F,CAAC;AAED,sFAAsF;AACtF,MAAM,UAAU,OAAO,CAAC,KAAa,EAAE,aAA4C;IACjF,IAAI,aAAa,KAAK,OAAO;QAAE,OAAO,QAAQ,CAAC,KAAK,CAAC,CAAC;IACtD,IAAI,aAAa,KAAK,QAAQ,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,GAAG,CAAC;QAAE,OAAO,GAAG,KAAK,GAAG,CAAC;IAC3E,OAAO,KAAK,CAAC;AACf,CAAC;AAmBD;;;;;GAKG;AACH,MAAM,UAAU,gBAAgB,CAAC,OAAuB;IACtD,MAAM,QAAQ,GAAG,IAAI,GAAG,CAAC,OAAO,CAAC,QAAQ,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC,CAAC;IACzD,MAAM,OAAO,GAAG,IAAI,GAAG,EAAU,CAAC;IAClC,MAAM,MAAM,GAAG,IAAI,GAAG,EAAkB,CAAC;IACzC,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,QAAQ,EAAE,CAAC;QACrC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,MAAM,GAAG,GAAG,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACjC,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YACjB,IAAI,IAAI,CAAC,OAAO,KAAK,UAAU,IAAI,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC;gBAAE,SAAS;YAC9F,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC;gBAAE,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC;QACpD,CAAC;IACH,CAAC;IACD,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,SAAS,EAAE,CAAC;QACtC,MAAM,GAAG,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC;QAC5B,IAAI,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC;YAAE,SAAS;QACvE,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;IACzB,CAAC;IACD,OAAO;QACL,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,GAAG,CAAC,KAAK,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC;KACnF,CAAC,IAAI,EAAE,CAAC;AACX,CAAC;AAED,SAAS,SAAS,CAAC,IAAY;IAC7B,OAAO,IAAI;SACR,UAAU,CAAC,GAAG,EAAE,OAAO,CAAC;SACxB,UAAU,CAAC,GAAG,EAAE,MAAM,CAAC;SACvB,UAAU,CAAC,GAAG,EAAE,MAAM,CAAC;SACvB,UAAU,CAAC,GAAG,EAAE,QAAQ,CAAC;SACzB,UAAU,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC;AAC/B,CAAC;AAED,+CAA+C;AAC/C,MAAM,UAAU,UAAU,CAAC,SAA4B;IACrD,MAAM,IAAI,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,eAAe,SAAS,CAAC,OAAO,CAAC,gBAAgB,CAAC,CAAC;IAC3F,OAAO,CACL,0CAA0C;QAC1C,gEAAgE;QAChE,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QACb,aAAa,CACd,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,GAAG,qBAAqB,CAAC;AAEzC;;;;;GAKG;AACH,MAAM,UAAU,UAAU,CACxB,QAAwC,EACxC,OAA2B;IAE3B,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,MAAM,MAAM,GAAe,EAAE,CAAC;IAC9B,KAAK,MAAM,KAAK,IAAI,QAAQ,EAAE,CAAC;QAC7B,MAAM,KAAK,GAAG,KAAK,CAAC,SAAS,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;QAC/D,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC;QACjE,IAAI,KAAK,KAAK,CAAC,CAAC;YAAE,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,KAAK,CAAa,CAAC,CAAC;IAChE,CAAC;IACD,MAAM,MAAM,GAAe,EAAE,CAAC;IAC9B,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC;QAAE,MAAM,CAAC,IAAI,CAAC,CAAC,eAAe,EAAE,GAAG,QAAQ,CAAC,CAAC,CAAC;IACrE,MAAM,CAAC,IAAI,CAAC,GAAG,MAAM,CAAC,CAAC;IACvB,IAAI,OAAO,KAAK,SAAS;QAAE,MAAM,CAAC,IAAI,CAAC,CAAC,YAAY,OAAO,EAAE,CAAC,CAAC,CAAC;IAChE,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC;IAC1C,OAAO,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC;AACrE,CAAC;AAED,qEAAqE;AACrE,MAAM,UAAU,YAAY,CAAC,IAAY;IACvC,OAAO,GAAG,aAAa,CAAC,IAAI,CAAC,aAAa,CAAC;AAC7C,CAAC","sourcesContent":["// The page records: `_ascribe/pages.json` in a build's site output\n// (output-layout contract §1.3), what each page tells every reader that isn't\n// a person. The integration writes from them a page's head (`Head.astro`),\n// the site's one sitemap, and its one `robots.txt`, from every project's\n// records together, at the host's root.\nimport { readFileSync, statSync } from \"node:fs\";\nimport path from \"node:path\";\nimport { normalizeBase } from \"./project.js\";\n\n/** The records' format version this reads. */\nconst VERSION = 1;\n\n/** How a page is treated: listed, listed and not to be quoted, or not listed. */\nexport type Listing = \"listed\" | \"unquoted\" | \"unlisted\";\n\n/** One page's record, with the route it's keyed by. */\nexport interface PageRecord {\n  /** The page's route, from the host's root. */\n  route: string;\n  /** The page's path from the content root, with `/`. */\n  source: string;\n  /** The page's title, as plain text. */\n  title?: string;\n  /** The page's description, as plain text on one line. */\n  description?: string;\n  /** The page's own address, absolute, when the project sets `[consumer] site`. */\n  canonical?: string;\n  listing: Listing;\n  /** The address of the page's Markdown version, from the host's root, when the build publishes one. */\n  markdown?: string;\n}\n\n/** `_ascribe/pages.json`. */\nexport interface PageRecords {\n  version: number;\n  site?: string;\n  pages: Record<string, Omit<PageRecord, \"route\">>;\n  robots: { lines: string[] };\n}\n\nfunction recordsFile(siteRoot: string): string {\n  return path.join(siteRoot, \"_ascribe\", \"pages.json\");\n}\n\n/** The records in a build's site output, or `undefined` when there aren't any. */\nexport function readRecords(siteRoot: string): PageRecords | undefined {\n  const file = recordsFile(siteRoot);\n  let text: string;\n  try {\n    text = readFileSync(file, \"utf8\");\n  } catch {\n    return undefined;\n  }\n  const records = JSON.parse(text) as PageRecords;\n  if (records.version !== VERSION) {\n    throw new Error(\n      `@ascribed/astro: ${file} is version ${String(records.version)} of the page records, and this version of the integration reads version ${VERSION}. Update @ascribed/astro and @ascribed/cli together.`,\n    );\n  }\n  return records;\n}\n\n/** A build's records by source path, as last read, and when its file was written. */\nconst bySource = new Map<string, { written: number; pages: Map<string, PageRecord> }>();\n\n/**\n * The record of the page whose source path is `source`, in the build whose\n * site output is at `siteRoot`, or `undefined` when it has none. Read again\n * when the file changes, as it does with each build in `astro dev`.\n */\nexport function recordFor(siteRoot: string, source: string): PageRecord | undefined {\n  let written: number;\n  try {\n    written = statSync(recordsFile(siteRoot)).mtimeMs;\n  } catch {\n    return undefined;\n  }\n  let known = bySource.get(siteRoot);\n  if (known === undefined || known.written !== written) {\n    const pages = new Map<string, PageRecord>();\n    for (const record of pageList(readRecords(siteRoot))) {\n      if (!pages.has(record.source)) pages.set(record.source, record);\n    }\n    known = { written, pages };\n    bySource.set(siteRoot, known);\n  }\n  return known.pages.get(source);\n}\n\n/** The records' pages, each with its route, in the file's order. */\nexport function pageList(records: PageRecords | undefined): PageRecord[] {\n  return Object.entries(records?.pages ?? {}).map(([route, page]) => ({ route, ...page }));\n}\n\n/**\n * A route as the sitemap compares it: with a leading `/` and no trailing one,\n * except at the root, so `/docs/` and `/docs` are one page.\n */\nexport function routeKey(route: string): string {\n  const withSlash = route.startsWith(\"/\") ? route : `/${route}`;\n  return withSlash.length > 1 && withSlash.endsWith(\"/\") ? withSlash.slice(0, -1) : withSlash;\n}\n\n/** A route ending the way Astro's `trailingSlash` says, as a built page's address. */\nexport function routeAs(route: string, trailingSlash: \"always\" | \"never\" | \"ignore\"): string {\n  if (trailingSlash === \"never\") return routeKey(route);\n  if (trailingSlash === \"always\" && !route.endsWith(\"/\")) return `${route}/`;\n  return route;\n}\n\n/** What the sitemap is drawn from. */\nexport interface SitemapSources {\n  /** Every project's records. */\n  projects: readonly PageRecord[][];\n  /**\n   * Whether a project's page has a route in the site: was built, or, in\n   * `astro dev`, matches one of the site's routes.\n   */\n  routed: (route: string) => boolean;\n  /** The site's own pages, each by its route from the host's root. */\n  sitePages: readonly string[];\n  /** Routes from the host's root to leave out: forwarding pages, the not-found page, and the site's list. */\n  excluded: readonly string[];\n  /** Astro's `site`. */\n  site: string;\n}\n\n/**\n * The sitemap's addresses, absolute, sorted, each once: every listed page of\n * every project that has a route in the site, and every page of the site's\n * own, less the excluded routes. A project's page that isn't listed is left\n * out, and so is a page of the site's own at the same route as one.\n */\nexport function sitemapAddresses(sources: SitemapSources): string[] {\n  const excluded = new Set(sources.excluded.map(routeKey));\n  const records = new Set<string>();\n  const routes = new Map<string, string>();\n  for (const pages of sources.projects) {\n    for (const page of pages) {\n      const key = routeKey(page.route);\n      records.add(key);\n      if (page.listing === \"unlisted\" || excluded.has(key) || !sources.routed(page.route)) continue;\n      if (!routes.has(key)) routes.set(key, page.route);\n    }\n  }\n  for (const route of sources.sitePages) {\n    const key = routeKey(route);\n    if (records.has(key) || excluded.has(key) || routes.has(key)) continue;\n    routes.set(key, route);\n  }\n  return [\n    ...new Set([...routes.values()].map((route) => new URL(route, sources.site).href)),\n  ].sort();\n}\n\nfunction escapeXml(text: string): string {\n  return text\n    .replaceAll(\"&\", \"&amp;\")\n    .replaceAll(\"<\", \"&lt;\")\n    .replaceAll(\">\", \"&gt;\")\n    .replaceAll('\"', \"&quot;\")\n    .replaceAll(\"'\", \"&apos;\");\n}\n\n/** A sitemap of `addresses`, with no dates. */\nexport function sitemapXml(addresses: readonly string[]): string {\n  const urls = addresses.map((address) => `  <url><loc>${escapeXml(address)}</loc></url>\\n`);\n  return (\n    '<?xml version=\"1.0\" encoding=\"UTF-8\"?>\\n' +\n    '<urlset xmlns=\"http://www.sitemaps.org/schemas/sitemap/0.9\">\\n' +\n    urls.join(\"\") +\n    \"</urlset>\\n\"\n  );\n}\n\nconst USER_AGENT = /^\\s*user-agent\\s*:/i;\n\n/**\n * A site's `robots.txt` from each project's lines: a `User-agent: *` group\n * with every project's lines before its first `User-agent` line, which are\n * for every crawler, then the groups each project wrote whole, then the\n * sitemap's address. `undefined` when it would say nothing.\n */\nexport function robotsText(\n  projects: readonly (readonly string[])[],\n  sitemap: string | undefined,\n): string | undefined {\n  const everyone: string[] = [];\n  const groups: string[][] = [];\n  for (const lines of projects) {\n    const first = lines.findIndex((line) => USER_AGENT.test(line));\n    everyone.push(...(first === -1 ? lines : lines.slice(0, first)));\n    if (first !== -1) groups.push(lines.slice(first) as string[]);\n  }\n  const blocks: string[][] = [];\n  if (everyone.length > 0) blocks.push([\"User-agent: *\", ...everyone]);\n  blocks.push(...groups);\n  if (sitemap !== undefined) blocks.push([`Sitemap: ${sitemap}`]);\n  if (blocks.length === 0) return undefined;\n  return `${blocks.map((block) => block.join(\"\\n\")).join(\"\\n\\n\")}\\n`;\n}\n\n/** Where the sitemap is, from the host's root: at Astro's `base`. */\nexport function sitemapRoute(base: string): string {\n  return `${normalizeBase(base)}sitemap.xml`;\n}\n"]}
+{"version":3,"file":"records.js","sourceRoot":"","sources":["../src/records.ts"],"names":[],"mappings":"AAAA,mEAAmE;AACnE,8EAA8E;AAC9E,2EAA2E;AAC3E,yEAAyE;AACzE,wCAAwC;AACxC,OAAO,EAAE,YAAY,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAC;AACjD,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,EAAE,aAAa,EAAE,MAAM,cAAc,CAAC;AAE7C,8CAA8C;AAC9C,MAAM,OAAO,GAAG,CAAC,CAAC;AAsFlB,SAAS,WAAW,CAAC,QAAgB;IACnC,OAAO,IAAI,CAAC,IAAI,CAAC,QAAQ,EAAE,UAAU,EAAE,YAAY,CAAC,CAAC;AACvD,CAAC;AAED,kFAAkF;AAClF,MAAM,UAAU,WAAW,CAAC,QAAgB;IAC1C,MAAM,IAAI,GAAG,WAAW,CAAC,QAAQ,CAAC,CAAC;IACnC,IAAI,IAAY,CAAC;IACjB,IAAI,CAAC;QACH,IAAI,GAAG,YAAY,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;IACpC,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,MAAM,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAgB,CAAC;IAChD,IAAI,OAAO,CAAC,OAAO,KAAK,OAAO,EAAE,CAAC;QAChC,MAAM,IAAI,KAAK,CACb,oBAAoB,IAAI,eAAe,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,2EAA2E,OAAO,sDAAsD,CACvM,CAAC;IACJ,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,qFAAqF;AACrF,MAAM,QAAQ,GAAG,IAAI,GAAG,EAA+D,CAAC;AAExF;;;;GAIG;AACH,MAAM,UAAU,SAAS,CAAC,QAAgB,EAAE,MAAc;IACxD,IAAI,OAAe,CAAC;IACpB,IAAI,CAAC;QACH,OAAO,GAAG,QAAQ,CAAC,WAAW,CAAC,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC;IACpD,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,IAAI,KAAK,GAAG,QAAQ,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;IACnC,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,CAAC,OAAO,KAAK,OAAO,EAAE,CAAC;QACrD,MAAM,KAAK,GAAG,IAAI,GAAG,EAAsB,CAAC;QAC5C,KAAK,MAAM,MAAM,IAAI,QAAQ,CAAC,WAAW,CAAC,QAAQ,CAAC,CAAC,EAAE,CAAC;YACrD,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,MAAM,CAAC;gBAAE,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;QAClE,CAAC;QACD,KAAK,GAAG,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC;QAC3B,QAAQ,CAAC,GAAG,CAAC,QAAQ,EAAE,KAAK,CAAC,CAAC;IAChC,CAAC;IACD,OAAO,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;AACjC,CAAC;AAED,oEAAoE;AACpE,MAAM,UAAU,QAAQ,CAAC,OAAgC;IACvD,OAAO,MAAM,CAAC,OAAO,CAAC,OAAO,EAAE,KAAK,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,KAAK,EAAE,IAAI,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,GAAG,IAAI,EAAE,CAAC,CAAC,CAAC;AAC3F,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,QAAQ,CAAC,KAAa;IACpC,MAAM,SAAS,GAAG,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,EAAE,CAAC;IAC9D,OAAO,SAAS,CAAC,MAAM,GAAG,CAAC,IAAI,SAAS,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;AAC9F,CAAC;AAED,sFAAsF;AACtF,MAAM,UAAU,OAAO,CAAC,KAAa,EAAE,aAA4C;IACjF,IAAI,aAAa,KAAK,OAAO;QAAE,OAAO,QAAQ,CAAC,KAAK,CAAC,CAAC;IACtD,IAAI,aAAa,KAAK,QAAQ,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,GAAG,CAAC;QAAE,OAAO,GAAG,KAAK,GAAG,CAAC;IAC3E,OAAO,KAAK,CAAC;AACf,CAAC;AAmBD;;;;;GAKG;AACH,MAAM,UAAU,gBAAgB,CAAC,OAAuB;IACtD,MAAM,QAAQ,GAAG,IAAI,GAAG,CAAC,OAAO,CAAC,QAAQ,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC,CAAC;IACzD,MAAM,OAAO,GAAG,IAAI,GAAG,EAAU,CAAC;IAClC,MAAM,MAAM,GAAG,IAAI,GAAG,EAAkB,CAAC;IACzC,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,QAAQ,EAAE,CAAC;QACrC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,MAAM,GAAG,GAAG,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACjC,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YACjB,IAAI,IAAI,CAAC,OAAO,KAAK,UAAU,IAAI,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC;gBAAE,SAAS;YAC9F,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC;gBAAE,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC;QACpD,CAAC;IACH,CAAC;IACD,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,SAAS,EAAE,CAAC;QACtC,MAAM,GAAG,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC;QAC5B,IAAI,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC;YAAE,SAAS;QACvE,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;IACzB,CAAC;IACD,OAAO;QACL,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,GAAG,CAAC,KAAK,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC;KACnF,CAAC,IAAI,EAAE,CAAC;AACX,CAAC;AAED,SAAS,SAAS,CAAC,IAAY;IAC7B,OAAO,IAAI;SACR,UAAU,CAAC,GAAG,EAAE,OAAO,CAAC;SACxB,UAAU,CAAC,GAAG,EAAE,MAAM,CAAC;SACvB,UAAU,CAAC,GAAG,EAAE,MAAM,CAAC;SACvB,UAAU,CAAC,GAAG,EAAE,QAAQ,CAAC;SACzB,UAAU,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC;AAC/B,CAAC;AAED,+CAA+C;AAC/C,MAAM,UAAU,UAAU,CAAC,SAA4B;IACrD,MAAM,IAAI,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,eAAe,SAAS,CAAC,OAAO,CAAC,gBAAgB,CAAC,CAAC;IAC3F,OAAO,CACL,0CAA0C;QAC1C,gEAAgE;QAChE,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QACb,aAAa,CACd,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,GAAG,qBAAqB,CAAC;AAEzC;;;;;;;GAOG;AACH,MAAM,UAAU,UAAU,CACxB,QAA2B,EAC3B,OAA2B;IAE3B,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAqB,CAAC;IAC9C,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,MAAM,MAAM,GAAe,EAAE,CAAC;IAC9B,KAAK,MAAM,MAAM,IAAI,QAAQ,EAAE,CAAC;QAC9B,KAAK,MAAM,KAAK,IAAI,MAAM,CAAC,MAAM,IAAI,EAAE,EAAE,CAAC;YACxC,IAAI,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC;gBAAE,SAAS;YAC1C,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,EAAE;gBAC1B,GAAG,KAAK,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,eAAe,KAAK,EAAE,CAAC;gBAC3D,GAAG,KAAK,CAAC,KAAK;aACf,CAAC,CAAC;QACL,CAAC;QACD,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC;QAC3B,MAAM,KAAK,GAAG,KAAK,CAAC,SAAS,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;QAC/D,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC;QACjE,IAAI,KAAK,KAAK,CAAC,CAAC;YAAE,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC;IACpD,CAAC;IACD,MAAM,MAAM,GAAe,CAAC,GAAG,QAAQ,CAAC,MAAM,EAAE,CAAC,CAAC;IAClD,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC;QAAE,MAAM,CAAC,IAAI,CAAC,CAAC,eAAe,EAAE,GAAG,QAAQ,CAAC,CAAC,CAAC;IACrE,MAAM,CAAC,IAAI,CAAC,GAAG,MAAM,CAAC,CAAC;IACvB,IAAI,OAAO,KAAK,SAAS;QAAE,MAAM,CAAC,IAAI,CAAC,CAAC,YAAY,OAAO,EAAE,CAAC,CAAC,CAAC;IAChE,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC;IAC1C,OAAO,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC;AACrE,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,gBAAgB,CAAC,QAAkC;IACjE,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,KAAK,MAAM,OAAO,IAAI,CAAC,UAAU,EAAE,QAAQ,EAAE,YAAY,CAAU,EAAE,CAAC;QACpE,MAAM,KAAK,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,QAAQ,EAAE,CAAC,OAAO,CAAC,KAAK,OAAO,CAAC,CAAC;QAC7E,MAAM,QAAQ,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,QAAQ,EAAE,CAAC,OAAO,CAAC,KAAK,UAAU,CAAC,CAAC;QACnF,IAAI,KAAK,KAAK,SAAS,IAAI,QAAQ,KAAK,SAAS;YAAE,SAAS;QAC5D,QAAQ,CAAC,IAAI,CACX,IAAI,QAAQ,CAAC,IAAI,oBAAoB,OAAO,sBAAsB,KAAK,CAAC,IAAI,SAAS,OAAO,yGAAyG,CACtM,CAAC;IACJ,CAAC;IACD,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,qEAAqE;AACrE,MAAM,UAAU,YAAY,CAAC,IAAY;IACvC,OAAO,GAAG,aAAa,CAAC,IAAI,CAAC,aAAa,CAAC;AAC7C,CAAC","sourcesContent":["// The page records: `_ascribe/pages.json` in a build's site output\n// (output-layout contract §1.3), what each page tells every reader that isn't\n// a person. The integration writes from them a page's head (`Head.astro`),\n// the site's one sitemap, and its one `robots.txt`, from every project's\n// records together, at the host's root.\nimport { readFileSync, statSync } from \"node:fs\";\nimport path from \"node:path\";\nimport { normalizeBase } from \"./project.js\";\n\n/** The records' format version this reads. */\nconst VERSION = 1;\n\n/** How a page is treated: listed, listed and not to be quoted, or not listed. */\nexport type Listing = \"listed\" | \"unquoted\" | \"unlisted\";\n\n/**\n * Where a page's listing comes from: `default` when nothing sets it, `page`\n * for the page's own, and for a rule of `[listing]` its level (`site`,\n * `folder`, or `type`), its folder or type, the rule as written, and its line\n * in `ascribe.toml`.\n */\nexport interface ListingFrom {\n  level: \"default\" | \"page\" | \"site\" | \"folder\" | \"type\";\n  folder?: string;\n  type?: string;\n  rule?: string;\n  line?: number;\n}\n\n/** One page's record, with the route it's keyed by. */\nexport interface PageRecord {\n  /** The page's route, from the host's root. */\n  route: string;\n  /** The page's path from the content root, with `/`. */\n  source: string;\n  /** The page's title, as plain text. */\n  title?: string;\n  /** The page's description, as plain text on one line. */\n  description?: string;\n  /** The page's own address, absolute, when the project sets `[consumer] site`. */\n  canonical?: string;\n  listing: Listing;\n  /** Where `listing` comes from: the page, a rule of `[listing]`, or neither. */\n  listing_from: ListingFrom;\n  /** The address of the page's Markdown version, from the host's root, when the build publishes one. */\n  markdown?: string;\n  /**\n   * The page's place in the site, from the top down to the page itself: the\n   * page at each of its route's folders that has one, then the page. Absent\n   * in records an earlier Ascribe wrote.\n   */\n  breadcrumbs?: Step[];\n}\n\n/** One step of a page's breadcrumbs: a page's route, from the host's root, and its title. */\nexport interface Step {\n  route: string;\n  title: string;\n}\n\n/** What a kind of crawler is for: gathering training data, building an AI search index, or fetching at a person's request. */\ntype Purpose = \"training\" | \"search\" | \"on-request\";\n\n/** The group of `robots.txt` a disallowed purpose writes. */\nexport interface RobotsGroup {\n  purpose: Purpose;\n  /** Every token of the purpose the binary knows, each a `User-agent` line. */\n  user_agents: string[];\n  /** The group's rules: `Disallow: /`. */\n  lines: string[];\n}\n\n/** What a project says to crawlers. */\nexport interface Robots {\n  /** The project's own lines, with paths from the host's root. */\n  lines: string[];\n  /** What the project says of each purpose it sets. */\n  purposes?: Partial<Record<Purpose, \"allow\" | \"disallow\">>;\n  /** A group for each purpose the project disallows. */\n  groups?: RobotsGroup[];\n}\n\n/** `_ascribe/pages.json`. */\nexport interface PageRecords {\n  version: number;\n  site?: string;\n  pages: Record<string, Omit<PageRecord, \"route\">>;\n  robots: Robots;\n}\n\n/** What one project says to crawlers, by the name of its first entry. */\nexport interface ProjectRobots {\n  name: string;\n  robots: Robots;\n}\n\nfunction recordsFile(siteRoot: string): string {\n  return path.join(siteRoot, \"_ascribe\", \"pages.json\");\n}\n\n/** The records in a build's site output, or `undefined` when there aren't any. */\nexport function readRecords(siteRoot: string): PageRecords | undefined {\n  const file = recordsFile(siteRoot);\n  let text: string;\n  try {\n    text = readFileSync(file, \"utf8\");\n  } catch {\n    return undefined;\n  }\n  const records = JSON.parse(text) as PageRecords;\n  if (records.version !== VERSION) {\n    throw new Error(\n      `@ascribed/astro: ${file} is version ${String(records.version)} of the page records, and this version of the integration reads version ${VERSION}. Update @ascribed/astro and @ascribed/cli together.`,\n    );\n  }\n  return records;\n}\n\n/** A build's records by source path, as last read, and when its file was written. */\nconst bySource = new Map<string, { written: number; pages: Map<string, PageRecord> }>();\n\n/**\n * The record of the page whose source path is `source`, in the build whose\n * site output is at `siteRoot`, or `undefined` when it has none. Read again\n * when the file changes, as it does with each build in `astro dev`.\n */\nexport function recordFor(siteRoot: string, source: string): PageRecord | undefined {\n  let written: number;\n  try {\n    written = statSync(recordsFile(siteRoot)).mtimeMs;\n  } catch {\n    return undefined;\n  }\n  let known = bySource.get(siteRoot);\n  if (known === undefined || known.written !== written) {\n    const pages = new Map<string, PageRecord>();\n    for (const record of pageList(readRecords(siteRoot))) {\n      if (!pages.has(record.source)) pages.set(record.source, record);\n    }\n    known = { written, pages };\n    bySource.set(siteRoot, known);\n  }\n  return known.pages.get(source);\n}\n\n/** The records' pages, each with its route, in the file's order. */\nexport function pageList(records: PageRecords | undefined): PageRecord[] {\n  return Object.entries(records?.pages ?? {}).map(([route, page]) => ({ route, ...page }));\n}\n\n/**\n * A route as the sitemap compares it: with a leading `/` and no trailing one,\n * except at the root, so `/docs/` and `/docs` are one page.\n */\nexport function routeKey(route: string): string {\n  const withSlash = route.startsWith(\"/\") ? route : `/${route}`;\n  return withSlash.length > 1 && withSlash.endsWith(\"/\") ? withSlash.slice(0, -1) : withSlash;\n}\n\n/** A route ending the way Astro's `trailingSlash` says, as a built page's address. */\nexport function routeAs(route: string, trailingSlash: \"always\" | \"never\" | \"ignore\"): string {\n  if (trailingSlash === \"never\") return routeKey(route);\n  if (trailingSlash === \"always\" && !route.endsWith(\"/\")) return `${route}/`;\n  return route;\n}\n\n/** What the sitemap is drawn from. */\nexport interface SitemapSources {\n  /** Every project's records. */\n  projects: readonly PageRecord[][];\n  /**\n   * Whether a project's page has a route in the site: was built, or, in\n   * `astro dev`, matches one of the site's routes.\n   */\n  routed: (route: string) => boolean;\n  /** The site's own pages, each by its route from the host's root. */\n  sitePages: readonly string[];\n  /** Routes from the host's root to leave out: forwarding pages, the not-found page, and the site's list. */\n  excluded: readonly string[];\n  /** Astro's `site`. */\n  site: string;\n}\n\n/**\n * The sitemap's addresses, absolute, sorted, each once: every listed page of\n * every project that has a route in the site, and every page of the site's\n * own, less the excluded routes. A project's page that isn't listed is left\n * out, and so is a page of the site's own at the same route as one.\n */\nexport function sitemapAddresses(sources: SitemapSources): string[] {\n  const excluded = new Set(sources.excluded.map(routeKey));\n  const records = new Set<string>();\n  const routes = new Map<string, string>();\n  for (const pages of sources.projects) {\n    for (const page of pages) {\n      const key = routeKey(page.route);\n      records.add(key);\n      if (page.listing === \"unlisted\" || excluded.has(key) || !sources.routed(page.route)) continue;\n      if (!routes.has(key)) routes.set(key, page.route);\n    }\n  }\n  for (const route of sources.sitePages) {\n    const key = routeKey(route);\n    if (records.has(key) || excluded.has(key) || routes.has(key)) continue;\n    routes.set(key, route);\n  }\n  return [\n    ...new Set([...routes.values()].map((route) => new URL(route, sources.site).href)),\n  ].sort();\n}\n\nfunction escapeXml(text: string): string {\n  return text\n    .replaceAll(\"&\", \"&amp;\")\n    .replaceAll(\"<\", \"&lt;\")\n    .replaceAll(\">\", \"&gt;\")\n    .replaceAll('\"', \"&quot;\")\n    .replaceAll(\"'\", \"&apos;\");\n}\n\n/** A sitemap of `addresses`, with no dates. */\nexport function sitemapXml(addresses: readonly string[]): string {\n  const urls = addresses.map((address) => `  <url><loc>${escapeXml(address)}</loc></url>\\n`);\n  return (\n    '<?xml version=\"1.0\" encoding=\"UTF-8\"?>\\n' +\n    '<urlset xmlns=\"http://www.sitemaps.org/schemas/sitemap/0.9\">\\n' +\n    urls.join(\"\") +\n    \"</urlset>\\n\"\n  );\n}\n\nconst USER_AGENT = /^\\s*user-agent\\s*:/i;\n\n/**\n * A site's `robots.txt` from what each project says to crawlers: a group for\n * each purpose a project disallows, once for the site, then a\n * `User-agent: *` group with every project's lines before its first\n * `User-agent` line, which are for every crawler, then the groups each\n * project wrote whole, then the sitemap's address. `undefined` when it would\n * say nothing.\n */\nexport function robotsText(\n  projects: readonly Robots[],\n  sitemap: string | undefined,\n): string | undefined {\n  const purposes = new Map<Purpose, string[]>();\n  const everyone: string[] = [];\n  const groups: string[][] = [];\n  for (const robots of projects) {\n    for (const group of robots.groups ?? []) {\n      if (purposes.has(group.purpose)) continue;\n      purposes.set(group.purpose, [\n        ...group.user_agents.map((agent) => `User-agent: ${agent}`),\n        ...group.lines,\n      ]);\n    }\n    const lines = robots.lines;\n    const first = lines.findIndex((line) => USER_AGENT.test(line));\n    everyone.push(...(first === -1 ? lines : lines.slice(0, first)));\n    if (first !== -1) groups.push(lines.slice(first));\n  }\n  const blocks: string[][] = [...purposes.values()];\n  if (everyone.length > 0) blocks.push([\"User-agent: *\", ...everyone]);\n  blocks.push(...groups);\n  if (sitemap !== undefined) blocks.push([`Sitemap: ${sitemap}`]);\n  if (blocks.length === 0) return undefined;\n  return `${blocks.map((block) => block.join(\"\\n\")).join(\"\\n\\n\")}\\n`;\n}\n\n/**\n * Two projects that say contradictory things about one purpose, as\n * sentences: one allows it and the other disallows it. A site has one\n * `robots.txt`, so they must agree. A project that says nothing of a purpose\n * agrees with either.\n */\nexport function purposeConflicts(projects: readonly ProjectRobots[]): string[] {\n  const problems: string[] = [];\n  for (const purpose of [\"training\", \"search\", \"on-request\"] as const) {\n    const allow = projects.find((p) => p.robots.purposes?.[purpose] === \"allow\");\n    const disallow = projects.find((p) => p.robots.purposes?.[purpose] === \"disallow\");\n    if (allow === undefined || disallow === undefined) continue;\n    problems.push(\n      `\"${disallow.name}\" has [crawlers] ${purpose} = \"disallow\" and \"${allow.name}\" has ${purpose} = \"allow\". The site has one robots.txt, so its projects must agree: make them the same, or remove one.`,\n    );\n  }\n  return problems;\n}\n\n/** Where the sitemap is, from the host's root: at Astro's `base`. */\nexport function sitemapRoute(base: string): string {\n  return `${normalizeBase(base)}sitemap.xml`;\n}\n"]}

package/dist/root-files.d.ts

@@ -1,9 +1,9 @@
 import type { IncomingMessage, ServerResponse } from "node:http";
-import { type SitemapSources } from "./records.js";
+import { type ProjectRobots, type SitemapSources } from "./records.js";
 /** What the files are drawn from. */
 export interface RootFileSources extends Omit<SitemapSources, "site"> {
-    /** Each project's `robots.txt` lines, from the host's root, each project once. */
-    lines: readonly (readonly string[])[];
+    /** What each project says to crawlers, each project once. */
+    robots: readonly ProjectRobots[];
     /** Astro's `site`, when it's set. */
     site: string | undefined;
     /** Astro's `base`. */
@@ -18,11 +18,16 @@ interface RootFileNote {
     level: "info" | "warn";
     text: string;
 }
-/** The files, each `undefined` when none is written, and what to say about them. */
+/**
+ * The files, each `undefined` when none is written, what to say about them,
+ * and the problems that stop `robots.txt` from being written, which fail a
+ * build.
+ */
 export interface RootFiles {
     sitemap: string | undefined;
     robots: string | undefined;
     notes: RootFileNote[];
+    problems: string[];
 }
 /** The sitemap and `robots.txt` for a site, and what the build says about them. */
 export declare function rootFiles(sources: RootFileSources): RootFiles;

package/dist/root-files.js

@@ -1,5 +1,5 @@
 import { normalizeBase } from "./project.js";
-import { robotsText, routeAs, routeKey, sitemapAddresses, sitemapRoute, sitemapXml, } from "./records.js";
+import { purposeConflicts, robotsText, routeAs, routeKey, sitemapAddresses, sitemapRoute, sitemapXml, } from "./records.js";
 /** The sitemap and `robots.txt` for a site, and what the build says about them. */
 export function rootFiles(sources) {
     const notes = [];
@@ -22,7 +22,10 @@ export function rootFiles(sources) {
         sitemap = sitemapXml(sitemapAddresses({ ...sources, site: sources.site }));
         sitemapAddress = new URL(sitemapRoute(base), sources.site).href;
     }
-    let robots = robotsText(sources.lines, sitemapAddress);
+    const problems = purposeConflicts(sources.robots);
+    let robots = problems.length > 0
+        ? undefined
+        : robotsText(sources.robots.map((project) => project.robots), sitemapAddress);
     if (robots !== undefined && base !== "/") {
         notes.push({
             level: "info",
@@ -37,7 +40,7 @@ export function rootFiles(sources) {
         });
         robots = undefined;
     }
-    return { sitemap, robots, notes };
+    return { sitemap, robots, notes, problems };
 }
 /**
  * A dev-server middleware serving the sitemap at `<base>sitemap.xml` and

package/dist/root-files.js.map

@@ -1 +1 @@
-{"version":3,"file":"root-files.js","sourceRoot":"","sources":["../src/root-files.ts"],"names":[],"mappings":"AASA,OAAO,EAAE,aAAa,EAAE,MAAM,cAAc,CAAC;AAC7C,OAAO,EACL,UAAU,EACV,OAAO,EACP,QAAQ,EACR,gBAAgB,EAChB,YAAY,EACZ,UAAU,GAEX,MAAM,cAAc,CAAC;AA6BtB,mFAAmF;AACnF,MAAM,UAAU,SAAS,CAAC,OAAwB;IAChD,MAAM,KAAK,GAAmB,EAAE,CAAC;IACjC,MAAM,IAAI,GAAG,aAAa,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;IACzC,IAAI,OAA2B,CAAC;IAChC,IAAI,cAAkC,CAAC;IACvC,IAAI,OAAO,CAAC,UAAU,EAAE,CAAC;QACvB,2FAA2F;QAC3F,KAAK,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,MAAM,EAAE,IAAI,EAAE,uDAAuD,EAAE,CAAC,CAAC;QAC7F,IAAI,OAAO,CAAC,IAAI,KAAK,SAAS;YAAE,cAAc,GAAG,IAAI,GAAG,CAAC,YAAY,CAAC,IAAI,CAAC,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC;IAClG,CAAC;SAAM,IAAI,OAAO,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QACtC,KAAK,CAAC,IAAI,CAAC;YACT,KAAK,EAAE,MAAM;YACb,IAAI,EAAE,yJAAyJ;SAChK,CAAC,CAAC;IACL,CAAC;SAAM,CAAC;QACN,OAAO,GAAG,UAAU,CAAC,gBAAgB,CAAC,EAAE,GAAG,OAAO,EAAE,IAAI,EAAE,OAAO,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;QAC3E,cAAc,GAAG,IAAI,GAAG,CAAC,YAAY,CAAC,IAAI,CAAC,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC;IAClE,CAAC;IACD,IAAI,MAAM,GAAG,UAAU,CAAC,OAAO,CAAC,KAAK,EAAE,cAAc,CAAC,CAAC;IACvD,IAAI,MAAM,KAAK,SAAS,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;QACzC,KAAK,CAAC,IAAI,CAAC;YACT,KAAK,EAAE,MAAM;YACb,IAAI,EAAE,uCAAuC,IAAI,kGAAkG,MAAM,EAAE;SAC5J,CAAC,CAAC;QACH,MAAM,GAAG,SAAS,CAAC;IACrB,CAAC;SAAM,IAAI,MAAM,KAAK,SAAS,IAAI,OAAO,CAAC,SAAS,EAAE,CAAC;QACrD,KAAK,CAAC,IAAI,CAAC;YACT,KAAK,EAAE,MAAM;YACb,IAAI,EAAE,gFAAgF,MAAM,EAAE;SAC/F,CAAC,CAAC;QACH,MAAM,GAAG,SAAS,CAAC;IACrB,CAAC;IACD,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC;AACpC,CAAC;AAID;;;;GAIG;AACH,MAAM,UAAU,mBAAmB,CAAC,IAAY,EAAE,KAAsB;IACtE,MAAM,WAAW,GAAG,YAAY,CAAC,IAAI,CAAC,CAAC;IACvC,OAAO,CAAC,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,EAAE;QACxB,IAAI,GAAG,CAAC,MAAM,KAAK,KAAK,IAAI,GAAG,CAAC,MAAM,KAAK,MAAM;YAAE,OAAO,IAAI,EAAE,CAAC;QACjE,oFAAoF;QACpF,MAAM,QAAQ,GACZ,CAAE,GAAgC,CAAC,WAAW,IAAI,GAAG,CAAC,GAAG,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QAC7F,IAAI,QAAQ,KAAK,WAAW,IAAI,QAAQ,KAAK,aAAa;YAAE,OAAO,IAAI,EAAE,CAAC;QAC1E,MAAM,OAAO,GAAG,KAAK,EAAE,CAAC;QACxB,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,GAChB,QAAQ,KAAK,WAAW;YACtB,CAAC,CAAC,CAAC,OAAO,CAAC,OAAO,EAAE,gCAAgC,CAAC;YACrD,CAAC,CAAC,CAAC,OAAO,CAAC,MAAM,EAAE,2BAA2B,CAAC,CAAC;QACpD,IAAI,IAAI,KAAK,SAAS;YAAE,OAAO,IAAI,EAAE,CAAC;QACtC,GAAG,CAAC,SAAS,CAAC,cAAc,EAAE,IAAI,CAAC,CAAC;QACpC,GAAG,CAAC,SAAS,CAAC,gBAAgB,EAAE,MAAM,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC;QACzD,GAAG,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,KAAK,MAAM,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;IACpD,CAAC,CAAC;AACJ,CAAC;AAiBD;;;;;GAKG;AACH,MAAM,UAAU,YAAY,CAAC,MAA6B,EAAE,KAAc;IACxE,MAAM,IAAI,GAAG,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IACvC,OAAO,MAAM;SACV,MAAM,CACL,CAAC,KAAK,EAAE,EAAE,CACR,KAAK,CAAC,IAAI,KAAK,MAAM;QACrB,KAAK,CAAC,MAAM,KAAK,SAAS;QAC1B,KAAK,CAAC,MAAM,CAAC,MAAM,KAAK,CAAC;QACzB,KAAK,CAAC,OAAO,KAAK,MAAM;QACxB,KAAK,CAAC,OAAO,KAAK,MAAM,CAC3B;SACA,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,OAAO,CAAC,GAAG,IAAI,GAAG,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,EAAE,EAAE,KAAK,CAAC,aAAa,CAAC,CAAC,CAAC;AAChG,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,SAAS,CACvB,MAA6B,EAC7B,KAAc;IAEd,MAAM,IAAI,GAAG,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IACvC,MAAM,KAAK,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,KAAK,MAAM,IAAI,KAAK,CAAC,MAAM,KAAK,SAAS,CAAC,CAAC;IAC5F,OAAO,CAAC,KAAK,EAAE,EAAE;QACf,IAAI,CAAC,GAAG,KAAK,GAAG,CAAC,UAAU,CAAC,IAAI,CAAC;YAAE,OAAO,KAAK,CAAC;QAChD,MAAM,KAAK,GAAG,IAAI,KAAK,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC;QAC7C,MAAM,KAAK,GAAG,IAAI,GAAG,CAAC,CAAC,KAAK,EAAE,QAAQ,CAAC,KAAK,CAAC,EAAE,GAAG,QAAQ,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC;QACvE,OAAO,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACvF,CAAC,CAAC;AACJ,CAAC","sourcesContent":["// The site's sitemap and `robots.txt`, from every project's page records\n// together: one of each for the site, at the host's root, which is above the\n// base path of every project that isn't mounted there. `astro build` writes\n// them into its output, and `astro dev` serves them.\n//\n// A site whose Astro `base` isn't `/` has no host root: its sitemap goes\n// under the base, and no `robots.txt` is written, since one under a path does\n// nothing. The build says what the host's own needs instead.\nimport type { IncomingMessage, ServerResponse } from \"node:http\";\nimport { normalizeBase } from \"./project.js\";\nimport {\n  robotsText,\n  routeAs,\n  routeKey,\n  sitemapAddresses,\n  sitemapRoute,\n  sitemapXml,\n  type SitemapSources,\n} from \"./records.js\";\n\n/** What the files are drawn from. */\nexport interface RootFileSources extends Omit<SitemapSources, \"site\"> {\n  /** Each project's `robots.txt` lines, from the host's root, each project once. */\n  lines: readonly (readonly string[])[];\n  /** Astro's `site`, when it's set. */\n  site: string | undefined;\n  /** Astro's `base`. */\n  base: string;\n  /** Whether the site has its own `robots.txt`, in `public/`. */\n  ownRobots: boolean;\n  /** Whether the site has its own `sitemap.xml`, in `public/`. */\n  ownSitemap: boolean;\n}\n\n/** A thing the build says about the files, and how loudly. */\ninterface RootFileNote {\n  level: \"info\" | \"warn\";\n  text: string;\n}\n\n/** The files, each `undefined` when none is written, and what to say about them. */\nexport interface RootFiles {\n  sitemap: string | undefined;\n  robots: string | undefined;\n  notes: RootFileNote[];\n}\n\n/** The sitemap and `robots.txt` for a site, and what the build says about them. */\nexport function rootFiles(sources: RootFileSources): RootFiles {\n  const notes: RootFileNote[] = [];\n  const base = normalizeBase(sources.base);\n  let sitemap: string | undefined;\n  let sitemapAddress: string | undefined;\n  if (sources.ownSitemap) {\n    // The site's own, at the same address, which robots.txt points to when there's an address.\n    notes.push({ level: \"info\", text: \"no sitemap.xml: public/sitemap.xml is the site's own.\" });\n    if (sources.site !== undefined) sitemapAddress = new URL(sitemapRoute(base), sources.site).href;\n  } else if (sources.site === undefined) {\n    notes.push({\n      level: \"warn\",\n      text: \"no sitemap.xml: a sitemap lists pages by their absolute addresses, and the site has no address. Set [consumer] site in ascribe.toml, or Astro's `site`.\",\n    });\n  } else {\n    sitemap = sitemapXml(sitemapAddresses({ ...sources, site: sources.site }));\n    sitemapAddress = new URL(sitemapRoute(base), sources.site).href;\n  }\n  let robots = robotsText(sources.lines, sitemapAddress);\n  if (robots !== undefined && base !== \"/\") {\n    notes.push({\n      level: \"info\",\n      text: `no robots.txt: Astro's \\`base\\` is \"${base}\", and a crawler reads robots.txt only at the host's root. Put these lines in the host's own:\\n${robots}`,\n    });\n    robots = undefined;\n  } else if (robots !== undefined && sources.ownRobots) {\n    notes.push({\n      level: \"info\",\n      text: `no robots.txt: public/robots.txt is the site's own. Ascribe's lines for it:\\n${robots}`,\n    });\n    robots = undefined;\n  }\n  return { sitemap, robots, notes };\n}\n\ntype Middleware = (req: IncomingMessage, res: ServerResponse, next: () => void) => void;\n\n/**\n * A dev-server middleware serving the sitemap at `<base>sitemap.xml` and\n * `robots.txt` at `/`, each as `files` has it when it's asked for, so each\n * rebuild's records are served.\n */\nexport function rootFilesMiddleware(base: string, files: () => RootFiles): Middleware {\n  const sitemapPath = sitemapRoute(base);\n  return (req, res, next) => {\n    if (req.method !== \"GET\" && req.method !== \"HEAD\") return next();\n    // Vite strips the base from `req.url`; connect keeps the original in `originalUrl`.\n    const pathname =\n      ((req as { originalUrl?: string }).originalUrl ?? req.url ?? \"\").split(/[?#]/, 1)[0] ?? \"\";\n    if (pathname !== sitemapPath && pathname !== \"/robots.txt\") return next();\n    const current = files();\n    const [text, type] =\n      pathname === sitemapPath\n        ? [current.sitemap, \"application/xml; charset=utf-8\"]\n        : [current.robots, \"text/plain; charset=utf-8\"];\n    if (text === undefined) return next();\n    res.setHeader(\"Content-Type\", type);\n    res.setHeader(\"Content-Length\", Buffer.byteLength(text));\n    res.end(req.method === \"HEAD\" ? undefined : text);\n  };\n}\n\n/** A route as `astro:routes:resolved` gives it, as much as these read. */\nexport interface AstroRoute {\n  pattern: string;\n  patternRegex: RegExp;\n  params: string[];\n  type: string;\n  origin: string;\n}\n\n/** Astro's routing, as much as these read. */\ninterface Routing {\n  base: string;\n  trailingSlash: \"always\" | \"never\" | \"ignore\";\n}\n\n/**\n * The site's own pages `astro dev` can name: each route of the site's own\n * with no parameters, other than the not-found and error pages, by its route\n * from the host's root. A page a route's `getStaticPaths` lists is known only\n * to a build.\n */\nexport function devSitePages(routes: readonly AstroRoute[], astro: Routing): string[] {\n  const root = normalizeBase(astro.base);\n  return routes\n    .filter(\n      (route) =>\n        route.type === \"page\" &&\n        route.origin === \"project\" &&\n        route.params.length === 0 &&\n        route.pattern !== \"/404\" &&\n        route.pattern !== \"/500\",\n    )\n    .map((route) => routeAs(`${root}${route.pattern.replace(/^\\//, \"\")}`, astro.trailingSlash));\n}\n\n/**\n * Whether a project's page, by its route from the host's root, has a route\n * in the site in `astro dev`: whether one of the site's own routes matches\n * it. A project no route shows, such as one whose collection a page only\n * quotes from, has no pages in the sitemap.\n */\nexport function devRouted(\n  routes: readonly AstroRoute[],\n  astro: Routing,\n): (route: string) => boolean {\n  const root = normalizeBase(astro.base);\n  const pages = routes.filter((route) => route.type === \"page\" && route.origin === \"project\");\n  return (route) => {\n    if (!`${route}/`.startsWith(root)) return false;\n    const under = `/${route.slice(root.length)}`;\n    const forms = new Set([under, routeKey(under), `${routeKey(under)}/`]);\n    return pages.some((page) => [...forms].some((form) => page.patternRegex.test(form)));\n  };\n}\n"]}
+{"version":3,"file":"root-files.js","sourceRoot":"","sources":["../src/root-files.ts"],"names":[],"mappings":"AASA,OAAO,EAAE,aAAa,EAAE,MAAM,cAAc,CAAC;AAC7C,OAAO,EACL,gBAAgB,EAChB,UAAU,EACV,OAAO,EACP,QAAQ,EACR,gBAAgB,EAChB,YAAY,EACZ,UAAU,GAGX,MAAM,cAAc,CAAC;AAkCtB,mFAAmF;AACnF,MAAM,UAAU,SAAS,CAAC,OAAwB;IAChD,MAAM,KAAK,GAAmB,EAAE,CAAC;IACjC,MAAM,IAAI,GAAG,aAAa,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;IACzC,IAAI,OAA2B,CAAC;IAChC,IAAI,cAAkC,CAAC;IACvC,IAAI,OAAO,CAAC,UAAU,EAAE,CAAC;QACvB,2FAA2F;QAC3F,KAAK,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,MAAM,EAAE,IAAI,EAAE,uDAAuD,EAAE,CAAC,CAAC;QAC7F,IAAI,OAAO,CAAC,IAAI,KAAK,SAAS;YAAE,cAAc,GAAG,IAAI,GAAG,CAAC,YAAY,CAAC,IAAI,CAAC,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC;IAClG,CAAC;SAAM,IAAI,OAAO,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QACtC,KAAK,CAAC,IAAI,CAAC;YACT,KAAK,EAAE,MAAM;YACb,IAAI,EAAE,yJAAyJ;SAChK,CAAC,CAAC;IACL,CAAC;SAAM,CAAC;QACN,OAAO,GAAG,UAAU,CAAC,gBAAgB,CAAC,EAAE,GAAG,OAAO,EAAE,IAAI,EAAE,OAAO,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;QAC3E,cAAc,GAAG,IAAI,GAAG,CAAC,YAAY,CAAC,IAAI,CAAC,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC;IAClE,CAAC;IACD,MAAM,QAAQ,GAAG,gBAAgB,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;IAClD,IAAI,MAAM,GACR,QAAQ,CAAC,MAAM,GAAG,CAAC;QACjB,CAAC,CAAC,SAAS;QACX,CAAC,CAAC,UAAU,CACR,OAAO,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,MAAM,CAAC,EAC/C,cAAc,CACf,CAAC;IACR,IAAI,MAAM,KAAK,SAAS,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;QACzC,KAAK,CAAC,IAAI,CAAC;YACT,KAAK,EAAE,MAAM;YACb,IAAI,EAAE,uCAAuC,IAAI,kGAAkG,MAAM,EAAE;SAC5J,CAAC,CAAC;QACH,MAAM,GAAG,SAAS,CAAC;IACrB,CAAC;SAAM,IAAI,MAAM,KAAK,SAAS,IAAI,OAAO,CAAC,SAAS,EAAE,CAAC;QACrD,KAAK,CAAC,IAAI,CAAC;YACT,KAAK,EAAE,MAAM;YACb,IAAI,EAAE,gFAAgF,MAAM,EAAE;SAC/F,CAAC,CAAC;QACH,MAAM,GAAG,SAAS,CAAC;IACrB,CAAC;IACD,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,KAAK,EAAE,QAAQ,EAAE,CAAC;AAC9C,CAAC;AAID;;;;GAIG;AACH,MAAM,UAAU,mBAAmB,CAAC,IAAY,EAAE,KAAsB;IACtE,MAAM,WAAW,GAAG,YAAY,CAAC,IAAI,CAAC,CAAC;IACvC,OAAO,CAAC,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,EAAE;QACxB,IAAI,GAAG,CAAC,MAAM,KAAK,KAAK,IAAI,GAAG,CAAC,MAAM,KAAK,MAAM;YAAE,OAAO,IAAI,EAAE,CAAC;QACjE,oFAAoF;QACpF,MAAM,QAAQ,GACZ,CAAE,GAAgC,CAAC,WAAW,IAAI,GAAG,CAAC,GAAG,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QAC7F,IAAI,QAAQ,KAAK,WAAW,IAAI,QAAQ,KAAK,aAAa;YAAE,OAAO,IAAI,EAAE,CAAC;QAC1E,MAAM,OAAO,GAAG,KAAK,EAAE,CAAC;QACxB,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,GAChB,QAAQ,KAAK,WAAW;YACtB,CAAC,CAAC,CAAC,OAAO,CAAC,OAAO,EAAE,gCAAgC,CAAC;YACrD,CAAC,CAAC,CAAC,OAAO,CAAC,MAAM,EAAE,2BAA2B,CAAC,CAAC;QACpD,IAAI,IAAI,KAAK,SAAS;YAAE,OAAO,IAAI,EAAE,CAAC;QACtC,GAAG,CAAC,SAAS,CAAC,cAAc,EAAE,IAAI,CAAC,CAAC;QACpC,GAAG,CAAC,SAAS,CAAC,gBAAgB,EAAE,MAAM,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC;QACzD,GAAG,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,KAAK,MAAM,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;IACpD,CAAC,CAAC;AACJ,CAAC;AAiBD;;;;;GAKG;AACH,MAAM,UAAU,YAAY,CAAC,MAA6B,EAAE,KAAc;IACxE,MAAM,IAAI,GAAG,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IACvC,OAAO,MAAM;SACV,MAAM,CACL,CAAC,KAAK,EAAE,EAAE,CACR,KAAK,CAAC,IAAI,KAAK,MAAM;QACrB,KAAK,CAAC,MAAM,KAAK,SAAS;QAC1B,KAAK,CAAC,MAAM,CAAC,MAAM,KAAK,CAAC;QACzB,KAAK,CAAC,OAAO,KAAK,MAAM;QACxB,KAAK,CAAC,OAAO,KAAK,MAAM,CAC3B;SACA,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,OAAO,CAAC,GAAG,IAAI,GAAG,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,EAAE,EAAE,KAAK,CAAC,aAAa,CAAC,CAAC,CAAC;AAChG,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,SAAS,CACvB,MAA6B,EAC7B,KAAc;IAEd,MAAM,IAAI,GAAG,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IACvC,MAAM,KAAK,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,KAAK,MAAM,IAAI,KAAK,CAAC,MAAM,KAAK,SAAS,CAAC,CAAC;IAC5F,OAAO,CAAC,KAAK,EAAE,EAAE;QACf,IAAI,CAAC,GAAG,KAAK,GAAG,CAAC,UAAU,CAAC,IAAI,CAAC;YAAE,OAAO,KAAK,CAAC;QAChD,MAAM,KAAK,GAAG,IAAI,KAAK,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC;QAC7C,MAAM,KAAK,GAAG,IAAI,GAAG,CAAC,CAAC,KAAK,EAAE,QAAQ,CAAC,KAAK,CAAC,EAAE,GAAG,QAAQ,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC;QACvE,OAAO,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACvF,CAAC,CAAC;AACJ,CAAC","sourcesContent":["// The site's sitemap and `robots.txt`, from every project's page records\n// together: one of each for the site, at the host's root, which is above the\n// base path of every project that isn't mounted there. `astro build` writes\n// them into its output, and `astro dev` serves them.\n//\n// A site whose Astro `base` isn't `/` has no host root: its sitemap goes\n// under the base, and no `robots.txt` is written, since one under a path does\n// nothing. The build says what the host's own needs instead.\nimport type { IncomingMessage, ServerResponse } from \"node:http\";\nimport { normalizeBase } from \"./project.js\";\nimport {\n  purposeConflicts,\n  robotsText,\n  routeAs,\n  routeKey,\n  sitemapAddresses,\n  sitemapRoute,\n  sitemapXml,\n  type ProjectRobots,\n  type SitemapSources,\n} from \"./records.js\";\n\n/** What the files are drawn from. */\nexport interface RootFileSources extends Omit<SitemapSources, \"site\"> {\n  /** What each project says to crawlers, each project once. */\n  robots: readonly ProjectRobots[];\n  /** Astro's `site`, when it's set. */\n  site: string | undefined;\n  /** Astro's `base`. */\n  base: string;\n  /** Whether the site has its own `robots.txt`, in `public/`. */\n  ownRobots: boolean;\n  /** Whether the site has its own `sitemap.xml`, in `public/`. */\n  ownSitemap: boolean;\n}\n\n/** A thing the build says about the files, and how loudly. */\ninterface RootFileNote {\n  level: \"info\" | \"warn\";\n  text: string;\n}\n\n/**\n * The files, each `undefined` when none is written, what to say about them,\n * and the problems that stop `robots.txt` from being written, which fail a\n * build.\n */\nexport interface RootFiles {\n  sitemap: string | undefined;\n  robots: string | undefined;\n  notes: RootFileNote[];\n  problems: string[];\n}\n\n/** The sitemap and `robots.txt` for a site, and what the build says about them. */\nexport function rootFiles(sources: RootFileSources): RootFiles {\n  const notes: RootFileNote[] = [];\n  const base = normalizeBase(sources.base);\n  let sitemap: string | undefined;\n  let sitemapAddress: string | undefined;\n  if (sources.ownSitemap) {\n    // The site's own, at the same address, which robots.txt points to when there's an address.\n    notes.push({ level: \"info\", text: \"no sitemap.xml: public/sitemap.xml is the site's own.\" });\n    if (sources.site !== undefined) sitemapAddress = new URL(sitemapRoute(base), sources.site).href;\n  } else if (sources.site === undefined) {\n    notes.push({\n      level: \"warn\",\n      text: \"no sitemap.xml: a sitemap lists pages by their absolute addresses, and the site has no address. Set [consumer] site in ascribe.toml, or Astro's `site`.\",\n    });\n  } else {\n    sitemap = sitemapXml(sitemapAddresses({ ...sources, site: sources.site }));\n    sitemapAddress = new URL(sitemapRoute(base), sources.site).href;\n  }\n  const problems = purposeConflicts(sources.robots);\n  let robots =\n    problems.length > 0\n      ? undefined\n      : robotsText(\n          sources.robots.map((project) => project.robots),\n          sitemapAddress,\n        );\n  if (robots !== undefined && base !== \"/\") {\n    notes.push({\n      level: \"info\",\n      text: `no robots.txt: Astro's \\`base\\` is \"${base}\", and a crawler reads robots.txt only at the host's root. Put these lines in the host's own:\\n${robots}`,\n    });\n    robots = undefined;\n  } else if (robots !== undefined && sources.ownRobots) {\n    notes.push({\n      level: \"info\",\n      text: `no robots.txt: public/robots.txt is the site's own. Ascribe's lines for it:\\n${robots}`,\n    });\n    robots = undefined;\n  }\n  return { sitemap, robots, notes, problems };\n}\n\ntype Middleware = (req: IncomingMessage, res: ServerResponse, next: () => void) => void;\n\n/**\n * A dev-server middleware serving the sitemap at `<base>sitemap.xml` and\n * `robots.txt` at `/`, each as `files` has it when it's asked for, so each\n * rebuild's records are served.\n */\nexport function rootFilesMiddleware(base: string, files: () => RootFiles): Middleware {\n  const sitemapPath = sitemapRoute(base);\n  return (req, res, next) => {\n    if (req.method !== \"GET\" && req.method !== \"HEAD\") return next();\n    // Vite strips the base from `req.url`; connect keeps the original in `originalUrl`.\n    const pathname =\n      ((req as { originalUrl?: string }).originalUrl ?? req.url ?? \"\").split(/[?#]/, 1)[0] ?? \"\";\n    if (pathname !== sitemapPath && pathname !== \"/robots.txt\") return next();\n    const current = files();\n    const [text, type] =\n      pathname === sitemapPath\n        ? [current.sitemap, \"application/xml; charset=utf-8\"]\n        : [current.robots, \"text/plain; charset=utf-8\"];\n    if (text === undefined) return next();\n    res.setHeader(\"Content-Type\", type);\n    res.setHeader(\"Content-Length\", Buffer.byteLength(text));\n    res.end(req.method === \"HEAD\" ? undefined : text);\n  };\n}\n\n/** A route as `astro:routes:resolved` gives it, as much as these read. */\nexport interface AstroRoute {\n  pattern: string;\n  patternRegex: RegExp;\n  params: string[];\n  type: string;\n  origin: string;\n}\n\n/** Astro's routing, as much as these read. */\ninterface Routing {\n  base: string;\n  trailingSlash: \"always\" | \"never\" | \"ignore\";\n}\n\n/**\n * The site's own pages `astro dev` can name: each route of the site's own\n * with no parameters, other than the not-found and error pages, by its route\n * from the host's root. A page a route's `getStaticPaths` lists is known only\n * to a build.\n */\nexport function devSitePages(routes: readonly AstroRoute[], astro: Routing): string[] {\n  const root = normalizeBase(astro.base);\n  return routes\n    .filter(\n      (route) =>\n        route.type === \"page\" &&\n        route.origin === \"project\" &&\n        route.params.length === 0 &&\n        route.pattern !== \"/404\" &&\n        route.pattern !== \"/500\",\n    )\n    .map((route) => routeAs(`${root}${route.pattern.replace(/^\\//, \"\")}`, astro.trailingSlash));\n}\n\n/**\n * Whether a project's page, by its route from the host's root, has a route\n * in the site in `astro dev`: whether one of the site's own routes matches\n * it. A project no route shows, such as one whose collection a page only\n * quotes from, has no pages in the sitemap.\n */\nexport function devRouted(\n  routes: readonly AstroRoute[],\n  astro: Routing,\n): (route: string) => boolean {\n  const root = normalizeBase(astro.base);\n  const pages = routes.filter((route) => route.type === \"page\" && route.origin === \"project\");\n  return (route) => {\n    if (!`${route}/`.startsWith(root)) return false;\n    const under = `/${route.slice(root.length)}`;\n    const forms = new Set([under, routeKey(under), `${routeKey(under)}/`]);\n    return pages.some((page) => [...forms].some((form) => page.patternRegex.test(form)));\n  };\n}\n"]}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ascribed/astro",
-  "version": "0.2.1-next.16",
+  "version": "0.2.1-next.17",
   "description": "The Astro integration for Ascribe.",
   "type": "module",
   "license": "MPL-2.0",
@@ -54,9 +54,9 @@
     "node": ">=24"
   },
   "dependencies": {
-    "@ascribed/cli": "0.2.1-next.16",
-    "@ascribed/elements": "0.2.1-next.16",
-    "@ascribed/review": "0.2.1-next.16",
+    "@ascribed/cli": "0.2.1-next.17",
+    "@ascribed/elements": "0.2.1-next.17",
+    "@ascribed/review": "0.2.1-next.17",
     "github-slugger": "^2.0.0",
     "property-information": "^7.2.0",
     "smol-toml": "^1.9.0"
@@ -92,7 +92,7 @@
   "publishConfig": {
     "access": "public"
   },
-  "gitHead": "0077fa1e1fcd4ef67f9c6d48575c3f06f98adce8",
+  "gitHead": "605791982664a254a310026f05ebe69184943131",
   "scripts": {
     "build": "tsc -p tsconfig.build.json",
     "typecheck": "tsc --noEmit",

package/src/Head.astro

@@ -1,40 +1,58 @@
 ---
-// What a page's head says about it to every reader that isn't a person, from
-// its record in the build's `_ascribe/pages.json`: its `<title>`, its
-// description, its canonical link, and `<meta name="robots" content="noindex">`
-// when it isn't listed. Each comes from the record, as the sitemap and
-// `llms.txt` do, so none can disagree with them. A layout includes it in
-// `<head>`, given the page's collection entry:
+// Everything a page's head says about the page, from its record in the
+// build's `_ascribe/pages.json`: its `<title>` and description, its canonical
+// link, `<meta name="robots">` (`noindex` when it isn't listed, `nosnippet`
+// when it's listed and not to be quoted), a link to its Markdown version when
+// the build publishes one and the page is offered, the Open Graph tags a link
+// preview reads, and its breadcrumbs as structured data. Each comes from the
+// record, as the sitemap and `llms.txt` do, so none can disagree with them. A
+// layout includes it in `<head>`, given the page's collection entry:
 //
-//   <Head entry={entry} title={(page, record) => `${page} · Loom`} />
+//   <Head
+//     entry={entry}
+//     title={(page, record) => (record.route === "/" ? "Loom" : `${page} · Loom`)}
+//     name="Loom"
+//     card={{ src: "/social-card.png", alt: "Loom's docs", width: 1200, height: 630 }}
+//   />
 //
-// `title` is a function from the page's title and its record, which has its
-// route, to the whole title. A page with no record, such as an Astro page of
-// the site's own, gets nothing: it writes its own head.
-import { pageRecord, type PageRecord } from "@ascribed/astro/content";
+// What the site knows and the record doesn't is given by property: `title`,
+// a function from the page's title and its record, which has its route, to
+// the whole title; `name`, the site's name, for `og:site_name` and, on the
+// page at the host's root, the site-name structured data; and `card`, the
+// page's social card. The layout keeps the character set, the viewport,
+// icons, theme colors, and styles. A page with no record, such as an Astro
+// page of the site's own, gets nothing: it writes its own head.
+import { headTags, pageRecord, scriptJson, type Card, type PageRecord } from "@ascribed/astro/content";
 
 interface Props {
   /** The page's collection entry, or `undefined` for a page that has none. */
   entry?: { filePath?: string | undefined; data?: unknown } | undefined;
   /** The whole title, from the page's title and its record. Default: the page's title. */
   title?: (title: string, record: PageRecord) => string;
+  /** The site's name. */
+  name?: string;
+  /** The page's social card. Without one, no image tags. */
+  card?: Card;
 }
 
-const { entry, title } = Astro.props;
+const { entry, title, name, card } = Astro.props;
 const record = entry === undefined ? undefined : pageRecord(entry);
-const whole =
-  record?.title === undefined
-    ? undefined
-    : title === undefined
-      ? record.title
-      : title(record.title, record);
+const tags =
+  record === undefined
+    ? []
+    : headTags(record, { title, name, card, address: Astro.site ?? Astro.url });
 ---
 
-{whole !== undefined && <title>{whole}</title>}
 {
-  record?.description !== undefined && (
-    <meta name="description" content={record.description} />
+  tags.map((tag) =>
+    tag.tag === "title" ? (
+      <title>{tag.text}</title>
+    ) : tag.tag === "meta" ? (
+      <meta {...tag.attributes} />
+    ) : tag.tag === "link" ? (
+      <link {...tag.attributes} />
+    ) : (
+      <script is:inline type="application/ld+json" set:html={scriptJson(tag.json)} />
+    ),
   )
 }
-{record?.canonical !== undefined && <link rel="canonical" href={record.canonical} />}
-{record?.listing === "unlisted" && <meta name="robots" content="noindex" />}

package/starter/src/layouts/Docs.astro

@@ -3,7 +3,7 @@ import { getCollection, type CollectionEntry } from "astro:content";
 import type { ComponentProps } from "astro/types";
 import Availability from "@ascribed/astro/Availability.astro";
 import Elements from "@ascribed/astro/Elements.astro";
-// The page's title, description, canonical link, and robots rule, from its record.
+// Everything the head says about the page, from its record.
 import Head from "@ascribed/astro/Head.astro";
 import { sidebar } from "@ascribed/astro/content";
 import { nav } from "../nav.ts";