feat(docs-linter): add lint for local links (gnolang#2416)

<!-- please provide a detailed description of the changes made in this
pull request. -->

## Description

This PR improves the docs linter by adding a `os.Stat` check for local
files. It will notify the user if a link to a local file is broken by
trying to find the file in question on the filesystem.

<details><summary>Contributors' checklist...</summary>

- [x] Added new tests, or not needed, or not feasible
- [x] Provided an example (e.g. screenshot) to aid review or the PR is
self-explanatory
- [x] Updated the official documentation or not needed
- [x] No breaking changes were made, or a `BREAKING CHANGE: xxx` message
was included in the description
- [x] Added references to related issues and PRs
- [x] Provided any useful hints for running manual tests
- [x] Added new benchmarks to [generated
graphs](https://gnoland.github.io/benchmarks), if any. More info
[here](https://github.com/gnolang/gno/blob/master/.benchmarks/README.md).
</details>

---------

Co-authored-by: deelawn <dboltz03@gmail.com>

.github/workflows/docs-404-checker.yml → .github/workflows/docs-linter.yml

@@ -1,4 +1,4 @@
-name: "docs / 404 checker"
+name: "docs / lint"
 
 on:
   push:

docs/gno-infrastructure/validators/connect-to-existing-chain.md

@@ -19,7 +19,7 @@ In this tutorial, you will learn how to start a local Gno node and connect to an
 ## 1. Initialize the node directory
 
 To initialize a new Gno.land node working directory (configuration and secrets), make sure to
-follow [Step 1](./validators-setting-up-a-new-chain#1-generate-the-node-directory-secrets--config) from the
+follow [Step 1](./setting-up-a-new-chain.md#1-generate-the-node-directory-secrets--config) from the
 chain setup tutorial.
 
 ## 2. Obtain the `genesis.json` of the remote chain

docs/how-to-guides/connecting-from-go.md

@@ -50,7 +50,7 @@ The `gnoclient` package exposes a `Client` struct containing a `Signer` and
 `RPCClient` connector. `Client` exposes all available functionality for talking
 to a Gno.land chain.
 
-```go 
+```go
 type Client struct {
     Signer    Signer           // Signer for transaction authentication
     RPCClient rpcclient.Client // gnolang/gno/tm2/pkg/bft/rpc/client

docs/reference/stdlibs/std/testing.md

@@ -11,8 +11,8 @@ func TestSetOrigPkgAddr(addr Address)
 func TestSetOrigSend(sent, spent Coins)
 func TestIssueCoins(addr Address, coins Coins)
 func TestSetRealm(realm Realm)
-func NewUserRealm(address Address)
-func NewCodeRealm(pkgPath string)
+func NewUserRealm(address Address) Realm
+func NewCodeRealm(pkgPath string) Realm
 ```
 
 ---
@@ -132,7 +132,7 @@ userRealm := std.NewUserRealm(addr)
 ## NewCodeRealm
 
 ```go
-func NewCodeRealm(pkgPath string)
+func NewCodeRealm(pkgPath string) Realm
 ```
 
 Creates a new code realm for testing purposes.

misc/docs-linter/errors.go

@@ -3,9 +3,10 @@ package main
 import "errors"
 
 var (
-	errEmptyPath             = errors.New("you need to pass in a path to scan")
-	err404Link               = errors.New("link returned a 404")
-	errFound404Links         = errors.New("found links resulting in a 404 response status")
-	errFoundUnescapedJSXTags = errors.New("found unescaped JSX tags")
-	errFoundLintItems        = errors.New("found items that need linting")
+	errEmptyPath                  = errors.New("you need to pass in a path to scan")
+	err404Link                    = errors.New("link returned a 404")
+	errFound404Links              = errors.New("found links resulting in a 404 response status")
+	errFoundUnescapedJSXTags      = errors.New("found unescaped JSX tags")
+	errFoundUnreachableLocalLinks = errors.New("found local links that stat fails on")
+	errFoundLintItems             = errors.New("found items that need linting")
 )

misc/docs-linter/jsx.go

@@ -1,21 +1,25 @@
 package main
 
 import (
-	"context"
+	"bytes"
 	"fmt"
 	"regexp"
 	"strings"
 )
 
+var (
+	reCodeBlocks = regexp.MustCompile("(?s)```.*?```")
+	reInlineCode = regexp.MustCompile("`[^`]*`")
+)
+
+// extractJSX extracts JSX tags from given file content
 func extractJSX(fileContent []byte) []string {
 	text := string(fileContent)
 
 	// Remove code blocks
-	reCodeBlocks := regexp.MustCompile("(?s)```.*?```")
 	contentNoCodeBlocks := reCodeBlocks.ReplaceAllString(text, "")
 
 	// Remove inline code
-	reInlineCode := regexp.MustCompile("`[^`]*`")
 	contentNoInlineCode := reInlineCode.ReplaceAllString(contentNoCodeBlocks, "")
 
 	// Extract JSX/HTML elements
@@ -34,23 +38,25 @@ func extractJSX(fileContent []byte) []string {
 	return filteredMatches
 }
 
-func lintJSX(fileUrlMap map[string][]string, ctx context.Context) error {
-	found := false
-	for filePath, tags := range fileUrlMap {
-		filePath := filePath
+func lintJSX(filepathToJSX map[string][]string) (string, error) {
+	var (
+		found  bool
+		output bytes.Buffer
+	)
+	for filePath, tags := range filepathToJSX {
 		for _, tag := range tags {
 			if !found {
-				fmt.Println("Tags that need checking:")
+				output.WriteString("Tags that need checking:\n")
 				found = true
 			}
 
-			fmt.Printf(">>> %s (found in file: %s)\n", tag, filePath)
+			output.WriteString(fmt.Sprintf(">>> %s (found in file: %s)\n", tag, filePath))
 		}
 	}
 
 	if found {
-		return errFoundUnescapedJSXTags
+		return output.String(), errFoundUnescapedJSXTags
 	}
 
-	return nil
+	return "", nil
 }

misc/docs-linter/links.go

@@ -3,105 +3,91 @@ package main
 import (
 	"bufio"
 	"bytes"
-	"context"
 	"fmt"
-	"golang.org/x/sync/errgroup"
-	"io"
-	"mvdan.cc/xurls/v2"
-	"net/http"
+	"os"
+	"path/filepath"
+	"regexp"
 	"strings"
-	"sync"
 )
 
-// extractUrls extracts URLs from a file and maps them to the file
-func extractUrls(fileContent []byte) []string {
+// Valid start to an embedmd link
+const embedmd = `[embedmd]:# `
+
+// Regular expression to match markdown links
+var regex = regexp.MustCompile(`]\((\.\.?/.+?)\)`)
+
+// extractLocalLinks extracts links to local files from the given file content
+func extractLocalLinks(fileContent []byte) []string {
 	scanner := bufio.NewScanner(bytes.NewReader(fileContent))
-	urls := make([]string, 0)
+	links := make([]string, 0)
 
 	// Scan file line by line
 	for scanner.Scan() {
 		line := scanner.Text()
 
-		// Extract links
-		rxStrict := xurls.Strict()
-		url := rxStrict.FindString(line)
+		// Check for embedmd links
+		if embedmdPos := strings.Index(line, embedmd); embedmdPos != -1 {
+			link := line[embedmdPos+len(embedmd)+1:]
+
+			// Find closing parentheses
+			if closePar := strings.LastIndex(link, ")"); closePar != -1 {
+				link = link[:closePar]
+			}
+
+			// Remove space
+			if pos := strings.Index(link, " "); pos != -1 {
+				link = link[:pos]
+			}
 
-		// Check for empty links and skip them
-		if url == " " || len(url) == 0 {
+			// Add link to be checked
+			links = append(links, link)
 			continue
 		}
 
-		// Look for http & https only
-		if strings.HasPrefix(url, "http://") || strings.HasPrefix(url, "https://") {
-			// Ignore localhost
-			if !strings.Contains(url, "localhost") && !strings.Contains(url, "127.0.0.1") {
-				urls = append(urls, url)
+		// Find all matches
+		matches := regex.FindAllString(line, -1)
+
+		// Extract and print the local file links
+		for _, match := range matches {
+			// Remove ]( from the beginning and ) from end of link
+			match = match[2 : len(match)-1]
+
+			// Remove markdown headers in links
+			if pos := strings.Index(match, "#"); pos != -1 {
+				match = match[:pos]
 			}
+
+			links = append(links, match)
 		}
 	}
 
-	return urls
+	return links
 }
 
-func lintLinks(fileUrlMap map[string][]string, ctx context.Context) error {
-	// Filter links by prefix & ignore localhost
-	// Setup parallel checking for links
-	g, _ := errgroup.WithContext(ctx)
-
+func lintLocalLinks(filepathToLinks map[string][]string, docsPath string) (string, error) {
 	var (
-		lock         sync.Mutex
-		notFoundUrls []string
+		found  bool
+		output bytes.Buffer
 	)
 
-	for filePath, urls := range fileUrlMap {
-		filePath := filePath
-		for _, url := range urls {
-			url := url
-			g.Go(func() error {
-				if err := checkUrl(url); err != nil {
-					lock.Lock()
-					notFoundUrls = append(notFoundUrls, fmt.Sprintf(">>> %s (found in file: %s)", url, filePath))
-					lock.Unlock()
-				}
-
-				return nil
-			})
-		}
-	}
+	for filePath, links := range filepathToLinks {
+		for _, link := range links {
+			path := filepath.Join(docsPath, filepath.Dir(filePath), link)
 
-	if err := g.Wait(); err != nil {
-		return err
-	}
+			if _, err := os.Stat(path); err != nil {
+				if !found {
+					output.WriteString("Could not find files with the following paths:\n")
+					found = true
+				}
 
-	// Print out the URLs that returned a 404 along with the file names
-	if len(notFoundUrls) > 0 {
-		fmt.Println("Links that need checking:")
-		for _, result := range notFoundUrls {
-			fmt.Println(result)
+				output.WriteString(fmt.Sprintf(">>> %s (found in file: %s)\n", link, filePath))
+			}
 		}
-
-		return errFound404Links
 	}
 
-	return nil
-}
-
-// checkUrl checks if a URL is a 404
-func checkUrl(url string) error {
-	// Attempt to retrieve the HTTP header
-	resp, err := http.Get(url)
-	if err != nil || resp.StatusCode == http.StatusNotFound {
-		return err404Link
-	}
-
-	// Ensure the response body is closed properly
-	cleanup := func(Body io.ReadCloser) error {
-		if err := Body.Close(); err != nil {
-			return fmt.Errorf("could not close response properly: %w", err)
-		}
-
-		return nil
+	if found {
+		return output.String(), errFoundUnreachableLocalLinks
 	}
 
-	return cleanup(resp.Body)
+	return "", nil
 }