Add support for linguist-detectable and linguist-documentation.

modules/git/repo_attribute.go

@@ -291,7 +291,7 @@ func (repo *Repository) CheckAttributeReader(commitID string) (*CheckAttributeRe
  }
 
  checker := &CheckAttributeReader{
- Attributes: []string{"linguist-vendored", "linguist-generated", "linguist-language", "gitlab-language"},
+ Attributes: []string{"linguist-vendored", "linguist-generated", "linguist-language", "gitlab-language", "linguist-documentation", "linguist-detectable"},
  Repo: repo,
  IndexFile: indexFilename,
  WorkTree: worktree,

modules/git/repo_language_stats.go

@@ -6,13 +6,35 @@ package git
 import (
  "strings"
  "unicode"
+
+ "code.gitea.io/gitea/modules/optional"
 )
 
 const (
  fileSizeLimit int64 = 16 * 1024 // 16 KiB
  bigFileSize int64 = 1024 * 1024 // 1 MiB
 )
 
+// true if "set"/"true", false if "unset"/"false", none otherwise
+func linguistToBool(attr map[string]string, name string) optional.Option[bool] {
+ if value, has := attr[name]; has && value != "unspecified" {
+ switch value {
+ case "set", "true":
+ return optional.Some(true)
+ case "unset", "false":
+ return optional.Some(false)
+ }
+ }
+ return optional.None[bool]()
+}
+
+func linguistToString(attr map[string]string, name string) optional.Option[string] {
+ if value, has := attr[name]; has && value != "unspecified" {
+ return optional.Some(value)
+ }
+ return optional.None[string]()
+}
+
 // mergeLanguageStats mergers language names with different cases. The name with most upper case letters is used.
 func mergeLanguageStats(stats map[string]int64) map[string]int64 {
  names := map[string]struct {

modules/git/repo_language_stats_gogit.go

@@ -11,6 +11,7 @@ import (
  "strings"
 
  "code.gitea.io/gitea/modules/analyze"
+ "code.gitea.io/gitea/modules/optional"
 
  "github.com/go-enry/go-enry/v2"
  "github.com/go-git/go-git/v5"
@@ -57,25 +58,47 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
  return nil
  }
 
- notVendored := false
- notGenerated := false
+ isVendored := optional.None[bool]()
+ isGenerated := optional.None[bool]()
+ isDocumentation := optional.None[bool]()
+ isDetectable := optional.None[bool]()
 
  if checker != nil {
  attrs, err := checker.CheckPath(f.Name)
  if err == nil {
- if vendored, has := attrs["linguist-vendored"]; has {
- if vendored == "set" || vendored == "true" {
- return nil
- }
- notVendored = vendored == "false"
+ isVendored = linguistToBool(attrs, "linguist-vendored")
+ if isVendored.ValueOrDefault(false) {
+ return nil
+ }
+
+ isGenerated = linguistToBool(attrs, "linguist-generated")
+ if isGenerated.ValueOrDefault(false) {
+ return nil
  }
- if generated, has := attrs["linguist-generated"]; has {
- if generated == "set" || generated == "true" {
- return nil
+
+ isDocumentation = linguistToBool(attrs, "linguist-documentation")
+ if isDocumentation.ValueOrDefault(false) {
+ return nil
+ }
+
+ isDetectable = linguistToBool(attrs, "linguist-detectable")
+ if !isDetectable.ValueOrDefault(true) {
+ return nil
+ }
+
+ hasLanguage := linguistToString(attrs, "linguist-language")
+ if hasLanguage.Value() == "" {
+ hasLanguage = linguistToString(attrs, "gitlab-language")
+ if hasLanguage.Has() {
+ language := hasLanguage.Value()
+ if idx := strings.IndexByte(language, '?'); idx >= 0 {
+ hasLanguage = optional.Some(language[:idx])
+ }
  }
- notGenerated = generated == "false"
  }
- if language, has := attrs["linguist-language"]; has && language != "unspecified" && language != "" {
+ if hasLanguage.Value() != "" {
+ language := hasLanguage.Value()
+
  // group languages, such as Pug -> HTML; SCSS -> CSS
  group := enry.GetLanguageGroup(language)
  if len(group) != 0 {
@@ -85,28 +108,14 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
  // this language will always be added to the size
  sizes[language] += f.Size
  return nil
- } else if language, has := attrs["gitlab-language"]; has && language != "unspecified" && language != "" {
- // strip off a ? if present
- if idx := strings.IndexByte(language, '?'); idx >= 0 {
- language = language[:idx]
- }
- if len(language) != 0 {
- // group languages, such as Pug -> HTML; SCSS -> CSS
- group := enry.GetLanguageGroup(language)
- if len(group) != 0 {
- language = group
- }
-
- // this language will always be added to the size
- sizes[language] += f.Size
- return nil
- }
  }
  }
  }
 
- if (!notVendored && analyze.IsVendor(f.Name)) || enry.IsDotFile(f.Name) ||
- enry.IsDocumentation(f.Name) || enry.IsConfiguration(f.Name) {
+ if (!isVendored.Has() && analyze.IsVendor(f.Name)) ||
+ enry.IsDotFile(f.Name) ||
+ (!isDocumentation.Has() && enry.IsDocumentation(f.Name)) ||
+ enry.IsConfiguration(f.Name) {
  return nil
  }
 
@@ -115,12 +124,10 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
  if f.Size <= bigFileSize {
  content, _ = readFile(f, fileSizeLimit)
  }
- if !notGenerated && enry.IsGenerated(f.Name, content) {
+ if !isGenerated.Has() && enry.IsGenerated(f.Name, content) {
  return nil
  }
 
- // TODO: Use .gitattributes file for linguist overrides
-
  language := analyze.GetCodeLanguage(f.Name, content)
  if language == enry.OtherLanguage || language == "" {
  return nil
@@ -138,7 +145,7 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
  included = langtype == enry.Programming || langtype == enry.Markup
  includedLanguage[language] = included
  }
- if included {
+ if included || isDetectable.ValueOrDefault(false) {
  sizes[language] += f.Size
  } else if len(sizes) == 0 && (firstExcludedLanguage == "" || firstExcludedLanguage == language) {
  firstExcludedLanguage = language

modules/git/repo_language_stats_nogogit.go

@@ -14,6 +14,7 @@ import (
 
  "code.gitea.io/gitea/modules/analyze"
  "code.gitea.io/gitea/modules/log"
+ "code.gitea.io/gitea/modules/optional"
 
  "github.com/go-enry/go-enry/v2"
 )
@@ -90,25 +91,47 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
  continue
  }
 
- notVendored := false
- notGenerated := false
+ isVendored := optional.None[bool]()
+ isGenerated := optional.None[bool]()
+ isDocumentation := optional.None[bool]()
+ isDetectable := optional.None[bool]()
 
  if checker != nil {
  attrs, err := checker.CheckPath(f.Name())
  if err == nil {
- if vendored, has := attrs["linguist-vendored"]; has {
- if vendored == "set" || vendored == "true" {
- continue
- }
- notVendored = vendored == "false"
+ isVendored = linguistToBool(attrs, "linguist-vendored")
+ if isVendored.ValueOrDefault(false) {
+ continue
+ }
+
+ isGenerated = linguistToBool(attrs, "linguist-generated")
+ if isGenerated.ValueOrDefault(false) {
+ continue
  }
- if generated, has := attrs["linguist-generated"]; has {
- if generated == "set" || generated == "true" {
- continue
+
+ isDocumentation = linguistToBool(attrs, "linguist-documentation")
+ if isDocumentation.ValueOrDefault(false) {
+ continue
+ }
+
+ isDetectable = linguistToBool(attrs, "linguist-detectable")
+ if !isDetectable.ValueOrDefault(true) {
+ continue
+ }
+
+ hasLanguage := linguistToString(attrs, "linguist-language")
+ if hasLanguage.Value() == "" {
+ hasLanguage = linguistToString(attrs, "gitlab-language")
+ if hasLanguage.Has() {
+ language := hasLanguage.Value()
+ if idx := strings.IndexByte(language, '?'); idx >= 0 {
+ hasLanguage = optional.Some(language[:idx])
+ }
  }
- notGenerated = generated == "false"
  }
- if language, has := attrs["linguist-language"]; has && language != "unspecified" && language != "" {
+ if hasLanguage.Value() != "" {
+ language := hasLanguage.Value()
+
  // group languages, such as Pug -> HTML; SCSS -> CSS
  group := enry.GetLanguageGroup(language)
  if len(group) != 0 {
@@ -118,29 +141,14 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
  // this language will always be added to the size
  sizes[language] += f.Size()
  continue
- } else if language, has := attrs["gitlab-language"]; has && language != "unspecified" && language != "" {
- // strip off a ? if present
- if idx := strings.IndexByte(language, '?'); idx >= 0 {
- language = language[:idx]
- }
- if len(language) != 0 {
- // group languages, such as Pug -> HTML; SCSS -> CSS
- group := enry.GetLanguageGroup(language)
- if len(group) != 0 {
- language = group
- }
-
- // this language will always be added to the size
- sizes[language] += f.Size()
- continue
- }
  }
-
  }
  }
 
- if (!notVendored && analyze.IsVendor(f.Name())) || enry.IsDotFile(f.Name()) ||
- enry.IsDocumentation(f.Name()) || enry.IsConfiguration(f.Name()) {
+ if (!isVendored.Has() && analyze.IsVendor(f.Name())) ||
+ enry.IsDotFile(f.Name()) ||
+ (!isDocumentation.Has() && enry.IsDocumentation(f.Name())) ||
+ enry.IsConfiguration(f.Name()) {
  continue
  }
 
@@ -173,7 +181,7 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
  return nil, err
  }
  }
- if !notGenerated && enry.IsGenerated(f.Name(), content) {
+ if !isGenerated.Has() && enry.IsGenerated(f.Name(), content) {
  continue
  }
 
@@ -196,13 +204,12 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
  included = langType == enry.Programming || langType == enry.Markup
  includedLanguage[language] = included
  }
- if included {
+ if included || isDetectable.ValueOrDefault(false) {
  sizes[language] += f.Size()
  } else if len(sizes) == 0 && (firstExcludedLanguage == "" || firstExcludedLanguage == language) {
  firstExcludedLanguage = language
  firstExcludedLanguageSize += f.Size()
  }
- continue
  }
 
  // If there are no included languages add the first excluded language