feat: add whitelist char support to nonascii check (#100)

This commit brings support of whitelisted characters during repo healthcheck/non-ascii file check. Supported by an extra switch to `repo-health-checker`, `-whitelistedChars`. The argument takes a comma-separated list of non-ASCII characters and ignores them during repo healthcheck. Illegal cmdline input is logged by the logger.

Co-Authored-By: GitHub Copilot <noreply@microsoft.com>
<details>
<summary>Copilot Prompt</summary>
<br>
This is a repo for an online judge orchestrator system «JOJ3». Under `cmd/` lies a source directory for a Go command, `repo-health-checker`. You tell from its name that it checks the repo for stuff like repo size, commit message, non-ASCII character usage, etc. before sending the work to the actual judging and grading system.

Now, I want the non-ASCII character checking function of the repo health checker to be flexible - it shall accept a list of non-ASCII characters and deem them acceptable.

## Your task

- Accept this new cmdline arg. In `cmd/repo-health-checker/main.go`, accept a new command line flag `-whitelisted-chars`, which shall take exactly one string of comma-separated non-ASCII characters. This string shall be passed to the actual healthcheck package.
- Respect this list while scanning the files. In `pkg/healthcheck/nonascii.go`, function `getNonASCII()`, we utilize a bufio *Scanner* to scan through all files for non-ASCII characters. We would like the list of acceptable chars to be passed from the cmdline to here, and modify the scanner logic to actually accept the corresponding characters.
- Error handling and reporting. This command line arg, `-whitelisted-chars`, could be completely abscent; in which case, no characters shall be escaped by default. The comma-separated list passed to the command may contain ASCII characters or multiple characters that are not properly separated; in which case, ignore that element, and report the incident via the SLog logging framework used in this project.
- Test your work. Create new testcases under `examples/healthcheck/` to reflect this change. Reflect to `examples/healthcheck/asciifile/` to learn about how to configure the repo health checker. Integrate your work to the Go test framework such that it could be invoked by running `make test` at the terminal.
  - Note: Use `git init` to init your testcase directory and make a initial commit - this project, JOJ3, only runs in Git repos.

## Notes
- Directory structure. `cmd/` for invokable commands, `pkg/` for the actual logic, `internal` - something you don't need to worry about.
- JOJ3 vs. Health Check. `joj3` is a separate executable; in this session we are only working on the `repo-health-checker`.
- Extras. Make sure to read `README.md` and the directory structure before you go; also, create To-do before you execute your plan.
</details>

Reviewed-on: https://focs.ji.sjtu.edu.cn/git/JOJ/JOJ3/pulls/100
Reviewed-by: 张泊明518370910136 <bomingzh@sjtu.edu.cn>
Co-authored-by: Mack Wang <mac-wang@sjtu.edu.cn>
Co-committed-by: Mack Wang <mac-wang@sjtu.edu.cn>

.gitmodules

@@ -70,3 +70,7 @@
 	path = examples/cpplint/simple
 	url = ssh://git@focs.ji.sjtu.edu.cn:2222/JOJ/JOJ3-examples.git
 	branch = cpplint/simple
+[submodule "examples/healthcheck/whitelistedchars-success"]
+	path = examples/healthcheck/whitelistedchars-success
+	url = ssh://git@focs.ji.sjtu.edu.cn:2222/JOJ/JOJ3-examples.git
+	branch = healthcheck/whitelistedchars-success

cmd/repo-health-checker/main.go

@@ -45,6 +45,7 @@ var (
 	checkFileNameList string
 	checkFileSumList  string
 	metaFile          []string
+	whitelistedChars  string
 	allowedDomainList string
 	actorCsvPath      string
 	showVersion       *bool
@@ -57,6 +58,7 @@ func init() {
 	flag.Float64Var(&repoSize, "repoSize", 2, "maximum size of the repo in MiB")
 	flag.StringVar(&checkFileNameList, "checkFileNameList", "", "comma-separated list of files to check")
 	flag.StringVar(&checkFileSumList, "checkFileSumList", "", "comma-separated list of expected checksums")
+	flag.StringVar(&whitelistedChars, "whitelistedChars", "", "comma-separated list of non-ASCII characters allowed in files")
 	flag.StringVar(&allowedDomainList, "allowedDomainList", "sjtu.edu.cn", "comma-separated list of allowed domains for commit author email")
 	flag.StringVar(&actorCsvPath, "actorCsvPath", "/home/tt/.config/joj/students.csv", "path to actor csv file")
 	parseMultiValueFlag(&metaFile, "meta", "meta files to check")
@@ -74,12 +76,14 @@ func main() {
 		"repoSize", repoSize,
 		"checkFileNameList", checkFileNameList,
 		"checkFileSumList", checkFileSumList,
+		"whitelistedChars", whitelistedChars,
 		"meta", metaFile,
 	)
 	res := healthcheck.All(
 		rootDir,
 		checkFileNameList,
 		checkFileSumList,
+		whitelistedChars,
 		allowedDomainList,
 		actorCsvPath,
 		metaFile,

examples/healthcheck/whitelistedchars-success

@@ -0,0 +1 @@
+Subproject commit bb9bc06fd5753e7338e9b3230b2fc3e3ce971a05

pkg/healthcheck/all.go

@@ -12,7 +12,8 @@ type Result struct {
 }
 
 func All(
-	rootDir, checkFileNameList, checkFileSumList, allowedDomainList, actorCsvPath string,
+	rootDir, checkFileNameList, checkFileSumList, whitelistedChars,
+	allowedDomainList, actorCsvPath string,
 	metaFile []string, repoSize float64,
 ) (res Result) {
 	var err error
@@ -44,7 +45,7 @@ func All(
 	} else {
 		res.Msg += "### Meta File Check Passed\n"
 	}
-	err = NonASCIIFiles(rootDir)
+	err = NonASCIIFiles(rootDir, whitelistedChars)
 	if err != nil {
 		res.Msg += fmt.Sprintf("### Non-ASCII Characters File Check Failed:\n%s\n", err.Error())
 		res.Failed = true

pkg/healthcheck/nonascii.go

@@ -8,11 +8,47 @@ import (
 	"path/filepath"
 	"strings"
 	"unicode"
+	"unicode/utf8"
 
 	"github.com/go-git/go-git/v5"
 	"github.com/go-git/go-git/v5/plumbing/format/gitattributes"
 )
 
+// Read the list of comma-separated allowed characters from command line and convert it to a hashmap.
+func parseWhitelistedChars(csv string) map[rune]struct{} {
+	whitelist := make(map[rune]struct{})
+	if strings.TrimSpace(csv) == "" {
+		return whitelist
+	}
+
+	for _, raw := range strings.Split(csv, ",") {
+		elem := strings.TrimSpace(raw)
+		if elem == "" {
+			slog.Warn("ignoring invalid whitelisted-chars element", "element", raw, "reason", "empty element")
+			continue
+		}
+
+		if utf8.RuneCountInString(elem) != 1 {
+			slog.Warn("ignoring invalid whitelisted-chars element", "element", elem, "reason", "element must be exactly one character")
+			continue
+		}
+
+		ch, _ := utf8.DecodeRuneInString(elem)
+		if ch == utf8.RuneError {
+			slog.Warn("ignoring invalid whitelisted-chars element", "element", elem, "reason", "invalid utf-8 rune")
+			continue
+		}
+		if ch <= unicode.MaxASCII {
+			slog.Warn("ignoring invalid whitelisted-chars element", "element", elem, "reason", "ASCII characters are not allowed")
+			continue
+		}
+
+		whitelist[ch] = struct{}{}
+	}
+
+	return whitelist
+}
+
 // getSubmodulePathsFromGoGit uses the go-git library to open the repository
 // at the given root path and retrieve a list of all submodule paths.
 // It returns a set of submodule paths for efficient lookup.
@@ -48,7 +84,7 @@ func getSubmodulePathsFromGoGit(root string) (map[string]struct{}, error) {
 
 // getNonASCII retrieves a list of files in the specified root directory that contain non-ASCII characters.
 // It searches for non-ASCII characters in each file's content and returns a list of paths to files containing non-ASCII characters.
-func getNonASCII(root string) ([]string, error) {
+func getNonASCII(root string, whitelist map[rune]struct{}) ([]string, error) {
 	var nonASCII []string
 	gitattrExist := true
 	var matcher gitattributes.Matcher
@@ -113,6 +149,9 @@ func getNonASCII(root string) ([]string, error) {
 		for scanner.Scan() {
 			cont := true
 			for _, c := range scanner.Text() {
+				if _, ok := whitelist[c]; ok {
+					continue
+				}
 				if c > unicode.MaxASCII {
 					nonASCII = append(nonASCII, "\t"+path)
 					cont = false
@@ -132,8 +171,10 @@ func getNonASCII(root string) ([]string, error) {
 
 // NonASCIIFiles checks for non-ASCII characters in files within the specified root directory.
 // It prints a message with the paths to files containing non-ASCII characters, if any.
-func NonASCIIFiles(root string) error {
+// Additionally it accept a list of whitelisted characters that are allowed, repo-wide.
-	nonASCII, err := getNonASCII(root)
+func NonASCIIFiles(root, whitelistedChars string) error {
+	whitelist := parseWhitelistedChars(whitelistedChars)
+	nonASCII, err := getNonASCII(root, whitelist)
 	if err != nil {
 		slog.Error("getting non-ascii", "err", err)
 		return fmt.Errorf("error getting non-ascii: %w", err)