Press n or j to go to the next uncovered block, b, p or k for the previous block.
| 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 | 1x 1x 1x 1x 1x 1x 1x 1504x 1504x 1504x 5734x 5734x 5734x 1410x 4324x 4230x 1504x 1x 4230x 4230x 1x 1x 94x 94x 94x 1x 6x 6x 6x 6x 1x 6x 6x 156x 6x 6x 80x 6x 1x 5x 1x 3965x 3965x 3965x 3965x 3965x 3965x 3965x 3965x 3965x 6x 6x 88x 4x 88x 3960x 11x 11x 11x 11x 60x 11x 351x 11x 1x 1x 54x 810x 1x 12x 180x 1x 3x 45x 45x 3x 54x 54x 54x 1620x 3x 3x 45x 3x 3x 12x 12x 180x 1x 7930x 7930x 7930x 40166x 5786576x 7930x 1x 3965x 3965x 3965x | import path from 'path';
import fs from 'fs';
import matter from 'gray-matter';
import { defaultLocale } from '@/constants/site';
import { isSafeSlug } from './slug';
// Path to posts directory
const POST_PATH = path.join(process.cwd(), 'data/blog');
/**
* @description Recursively find all .mdx files in a directory
*/
const findMdxFiles = (dir: string): string[] => {
const files: string[] = [];
const items = fs.readdirSync(dir);
for (const item of items) {
const fullPath = path.join(dir, item);
const stat = fs.statSync(fullPath);
if (stat.isDirectory()) {
files.push(...findMdxFiles(fullPath));
} else if (item.endsWith('.mdx')) {
files.push(fullPath);
}
}
return files;
};
type ParsedPost = {
content: string;
fileName: string;
// eslint-disable-next-line @typescript-eslint/no-explicit-any
data: Record<string, any>;
};
/**
* @description Read and gray-matter-parse a single MDX file into its raw post shape.
* @param {string} filePath - Absolute path to the .mdx file.
* @returns {ParsedPost}
*/
const buildPost = (filePath: string): ParsedPost => {
const { data, content } = matter(fs.readFileSync(filePath, 'utf8'));
return { content, data, fileName: filePath.split('/').pop() ?? '' };
};
// The blog corpus is static per deploy, so read + parse every file ONCE and reuse it.
// Before this, `findPostBySlug` scanned and gray-matter-parsed the WHOLE corpus to resolve a
// single slug (O(N)); `getAllPosts` called it per slug (O(N²)); and `getAllCategories` called
// `getAllPosts` once per category/tag (O(C·N²)) — thousands of disk parses per page render.
// On-demand (fallback: 'blocking') tag URLs ran all of it inside a request-time serverless
// function, which blew past Vercel's timeout (FUNCTION_INVOCATION_TIMEOUT). Cache only in
// production so `next dev` still hot-reloads content edits.
let corpusCache: ParsedPost[] | null = null;
const loadCorpus = (): ParsedPost[] => {
Iif (corpusCache && process.env.NODE_ENV === 'production') {
return corpusCache;
}
corpusCache = findMdxFiles(POST_PATH).map(buildPost);
return corpusCache;
};
/**
* @description Reduce a slug or route params object to the bare, comparable slug.
* @param {string | { params: { slug: string } }} slug - Slug or route params.
* @returns {string} NFC-normalized slug.
* @throws {Error} When the slug looks like a directory traversal attempt.
*/
const normalizeSlug = (slug: string | { params: { slug: string } }): string => {
const input = typeof slug === 'object' ? slug.params.slug : slug;
const bare = input.split('/').pop() ?? input;
Iif (!isSafeSlug(bare)) {
throw new Error('Invalid slug parameter');
}
// Slugs with non-ASCII characters (e.g. "compoñentes") may arrive NFD-decomposed from some
// clients/crawlers while the frontmatter stores them NFC-composed.
return bare.normalize('NFC');
};
/**
* @description Find a parsed post by slug in the cached corpus, preferring the requested locale.
*
* SDD-L04. This used to match on slug alone, which had two consequences. The visible one was that
* every post answered under every locale prefix, and combined with a router-derived canonical it let
* Google index the same Spanish article twice. The subtler one: **two posts can legitimately share a
* slug across locales** — `github-workflows.es.mdx` and `.gl.mdx` both use
* `integracion-continua-con-github-actions-workflow`, because the phrase is spelled the same in
* Spanish and Galician — and a slug-only lookup returned whichever came first in the corpus, so
* `/gl/blog/nextjs/<slug>` served the *Spanish* post under a Galician prefix.
*
* Filtering by locale first fixes both. A slug that exists in no locale, or only in another, throws
* and the page 404s.
*
* @param {string | { params: { slug: string } }} slug - Slug or route params.
* @param {string} [locale] - Locale to resolve within. Omitted only by callers that genuinely want
* any match, of which there are none today.
* @returns {ParsedPost}
*/
const findPostBySlug = (slug: string | { params: { slug: string } }, locale?: string): ParsedPost => {
const normalizedSlug = normalizeSlug(slug);
const matchesSlug = ({ data, fileName }: ParsedPost) =>
data.slug?.normalize('NFC') === normalizedSlug || fileName.normalize('NFC') === `${normalizedSlug}.mdx`;
const corpus = loadCorpus();
const post = locale
? corpus.find((entry) => entry.data.locale === locale && matchesSlug(entry))
: corpus.find(matchesSlug);
if (!post) {
throw new Error('Post not found');
}
return post;
};
/**
* @description Extract the publication date from the `<Date date="MM-DD-YYYY" />`
* tag embedded in the post body. Returns an ISO date string (YYYY-MM-DD) or null,
* so it can be serialized in getStaticProps and used for JSON-LD datePublished.
*
* @param {string} content - Raw MDX content of the post.
* @returns {string | null} - ISO date string or null if not found/invalid.
*/
const extractPostDate = (content: string): string | null => {
const match = content.match(/<Date\s+date="([^"]+)"/);
Iif (!match) return null;
// Build the ISO string from the MM-DD-YYYY components directly: V8 parses non-ISO
// date strings as LOCAL midnight, so round-tripping through `new Date(...)` +
// `toISOString()` shifted every post one day early in timezones ahead of UTC
// (e.g. Europe/Madrid: "07-18-2026" → "2026-07-17").
const componentsMatch = match[1].match(/^(\d{2})-(\d{2})-(\d{4})$/);
Iif (!componentsMatch) return null;
const [, month, day, year] = componentsMatch;
// Date.UTC normalizes overflow ("13-40-2023" would roll over), so a round-trip
// mismatch means the components were not a real calendar date.
const utcDate = new Date(Date.UTC(Number(year), Number(month) - 1, Number(day)));
Iif (utcDate.getUTCMonth() !== Number(month) - 1 || utcDate.getUTCDate() !== Number(day)) {
return null;
}
// SDD-L04: return a full ISO instant, not a bare date. `2026-07-22` is valid ISO 8601 and Google
// accepts it, but its own structured-data guidance asks for a timezone so freshness is
// unambiguous rather than interpreted against the crawler's clock. `PostByline` already pins
// display formatting to `timeZone: 'UTC'` and the sitemap's <lastmod> accepts the full form, so
// nothing downstream has to change.
return `${year}-${month}-${day}T00:00:00+00:00`;
};
/**
* @description Get post by slug.
*
* @example
* getPostBySlug('first-post')
* returns [{ content: '...', meta: {...} }]
*
* @param {string} slug - Slug of the post.
* @returns {Object} - Object with post content and meta data.
*/
/**
* @description Shape a parsed corpus entry into the public { content, meta } post form.
* @param {ParsedPost} parsed - Raw parsed post from the corpus.
* @returns {Object} - Object with post content and derived meta.
*/
const toPost = ({ content, data }: ParsedPost) => ({
content,
meta: {
readTime: readingTime(content),
numberOfWords: countWords(content),
date: extractPostDate(content),
// Optional frontmatter field marking a substantive content update (freshness
// signal for search + LLM engines; feeds dateModified and sitemap lastmod)
updated: data.updated ?? null,
slug: data.slug,
title: data.title,
locale: data.locale,
category: data.category,
author: data.author,
tags: data.tags,
excerpt: data.excerpt,
image: data.image,
description: data.description,
alternate: data.alternate,
// null (not undefined) so the meta object survives getStaticProps serialization
faq: data.faq ?? null,
},
});
export const getPostBySlug = (slug: string | { params: { slug: string } }, locale?: string) =>
toPost(findPostBySlug(slug, locale));
/**
* @description Get all posts (read + parsed once, cached per deploy in production).
*
* @example
* getAllPosts();
* returns[
* {
* content: '...',
* meta: '...',
* }
* ];
*
* @returns {Object} - Object with posts.
*/
export const getAllPosts = () => loadCorpus().map(toPost);
/**
* @description Get all posts by locale.
*
* @example
* getPostsByLocale('en');
* returns[
* {
* content: '...',
* meta: '...',
* }
* ];
*
* @param {string} locale - Locale of the posts.
* @returns {Object} - Object with posts.
*/
export const getPostsByLocale = (locale: string) => {
const posts = getAllPosts();
return posts.filter((post) => post.meta.locale === locale);
};
/**
* @description Path of the newest post, used as the blog's landing URL: the Dock links to
* /blog and that route redirects here, so the entry point follows publishing instead of being
* pinned to a slug that goes stale (and 404s outright if the post is ever renamed or removed).
*
* Dates are the ISO strings extracted from the post body, so comparing them as text orders them
* chronologically. Scoping to a category serves /blog/<category>, which should land on that
* category's newest post rather than the blog's; an empty category falls back to the newest
* post overall, and a locale with no posts at all falls back to the default locale.
*
* @example
* getLatestPostPath('es');
* returns '/blog/nextjs/integracion-continua-con-github-actions-workflow'
*
* @param {string} locale - Locale of the posts.
* @param {string} [category] - Optional category to scope the search to.
* @returns {string | null} - Post path without locale prefix, or null when the corpus is empty.
*/
export const getLatestPostPath = (locale: string, category?: string): string | null => {
const localePosts = getPostsByLocale(locale);
const posts = localePosts.length ? localePosts : getPostsByLocale(defaultLocale);
const inCategory = category
? posts.filter((post) => post.meta.category.toLowerCase() === category.toLowerCase())
: [];
const pool = inCategory.length ? inCategory : posts;
const latest = [...pool].sort((a, b) => (b.meta.date ?? '').localeCompare(a.meta.date ?? ''))[0];
return latest ? `/blog/${latest.meta.category.toLowerCase()}/${latest.meta.slug}` : null;
};
/**
* @description Get all posts by category and locale.
*
* @example
* getPostsByLocaleAndCategory('en', 'JavaScript');
* returns[
* {
* content: '...',
* meta: '...',
* }
* ];
*
* @param {string} locale - Locale of the posts.
* @param {string} category - Category of the posts.
* @returns {Object} - Object with posts.
*/
export const getPostsByLocaleAndCategory = (locale: string, category: string) => {
const posts = getPostsByLocale(locale);
return posts.filter(
(post) => post.meta.category.toLowerCase() === category?.toLowerCase() || post.meta.tags.includes(category)
);
};
/**
* @description Get all posts by tag and locale.
*
* @example
* getPostsByTag('JavaScript', 'en');
* returns[
* {
* content: '...',
* meta: '...',
* }
* ];
*
* @param {string} tag - Tag of the posts.
* @param {string} locale - Locale of the posts.
* @returns {Object} - Object with posts.
*/
const getPostsByTag = (tag: string, locale: string) => {
const posts = getPostsByLocale(locale);
return posts.filter((post) => post.meta.tags.includes(tag));
};
/**
* @description Get all posts by category and locale.
*
* @example
* getPostsByCategory('JavaScript', 'en');
* returns[
* {
* content: '...',
* meta: '...',
* }
* ];
*
* @param {string} category - Category of the posts.
* @param {string} locale - Locale of the posts.
* @returns {Object} - Object with posts.
*/
const getPostsByCategory = (category: string, locale: string) => {
const posts = getPostsByLocale(locale);
return posts.filter((post) => post.meta.category === category);
};
/**
* @description Get all tags from all posts by locale.
*
* @example
* getAllTags('en');
* returns[
* {
* tag: 'JavaScript',
* total: 2,
* href: '/blog/javascript/first-post',
* }
* ];
*
* @param {string} locale - Locale of the posts.
* @returns {Object} - Object with tags.
*/
const getAllTags = (locale: string) => {
const posts = getPostsByLocale(locale);
const tags = posts.map((post) => post.meta.tags);
// Categories and tags share the /blog/[category]/ URL namespace, so a tag that spells a
// category name ('nextjs' vs category 'Nextjs') resolves to the SAME URL and shows up
// twice in the sidebar — once under Topics, once under Tags — with different counts.
// The SEO content pass (PR #131) introduced several of these. Frontmatter is cleaned, and
// this guard keeps the taxonomies disjoint if a colliding tag is ever added back.
const categoryNames = new Set(posts.map((post) => post.meta.category.toLowerCase()));
return [...new Set(tags.flat())]
.filter((tag) => !categoryNames.has(tag.toLowerCase()))
.map((tag) => {
const postsBytag = getPostsByTag(tag, locale);
return {
tag,
total: tags.flat().filter((t) => t === tag).length,
href: `/blog/${tag.toLowerCase()}/${postsBytag[0].meta.slug}`,
};
});
};
/**
* @description Get all categories from all posts by locale.
*
* @example
* getAllCategories('en')
* returns { categories: [], tags: [] }
*
* @param {string} locale - Locale of the posts.
* @returns {Object} - Object with categories.
*/
export const getAllCategories = (locale: string) => {
const posts = getPostsByLocale(locale);
const categories = posts.map((post) => post.meta.category);
const tags = getAllTags(locale);
return {
categories: [...new Set(categories)].map((category) => {
const postsByCategory = getPostsByCategory(category, locale);
return {
category,
total: categories.filter((c) => c === category).length,
href: `/blog/${category.toLowerCase()}/${postsByCategory[0].meta.slug}`,
};
}),
tags,
};
};
/**
* @description Count words in a string.
*
* @example
* countWords('Hello world')
* returns 2
*
* @param {string} content - String to count words.
* @returns {number} - Number of words.
*/
export const countWords = (content: string) => {
// Fixed: Use a more secure approach to avoid regex DoS
// Split by triple backticks and count words only in non-code sections
const parts = content.split('```');
let nonCodeContent = '';
// Take every other part (non-code sections)
for (let i = 0; i < parts.length; i += 2) {
nonCodeContent += parts[i] || '';
}
const nonCodeWords = nonCodeContent.split(/\s/g).filter((word) => word.length > 0);
return nonCodeWords.length;
};
/**
* @description Calculate reading time in minutes.
*
* @example
* readingTime('Hello world. This is a test. This is a test')
* returns 2
*
* @param {string} content - String to calculate reading time.
* @returns {number} - Reading time in minutes.
*/
export const readingTime = (content: string) => {
const wordsPerMinute = 200;
const numberOfWords = countWords(content);
return Math.ceil(numberOfWords / wordsPerMinute);
};
|