Improve file detection with signature check capabilities

internal/buffer/buffer.go

@@ -692,6 +692,16 @@ func (b *Buffer) UpdateRules() {
 	if ft == "off" {
 		return
 	}
+
+	// syntaxFileBuffer is a helper structure
+	// to store properties of one single syntax file
+	type syntaxFileBuffer struct {
+		header    *highlight.Header
+		fileName  string
+		syntaxDef *highlight.Def
+	}
+
+	syntaxFiles := []syntaxFileBuffer{}
 	syntaxFile := ""
 	foundDef := false
 	var header *highlight.Header
@@ -714,41 +724,79 @@ func (b *Buffer) UpdateRules() {
 			continue
 		}
 
-		if ((ft == "unknown" || ft == "") && highlight.MatchFiletype(header.FtDetect, b.Path, b.lines[0].data)) || header.FileType == ft {
+		if ((ft == "unknown" || ft == "") && header.MatchFileName(b.Path)) || header.FileType == ft {
 			syndef, err := highlight.ParseDef(file, header)
 			if err != nil {
 				screen.TermMessage("Error parsing syntax file " + f.Name() + ": " + err.Error())
 				continue
 			}
-			b.SyntaxDef = syndef
-			syntaxFile = f.Name()
 			foundDef = true
-			break
+
+			if header.FileType == ft {
+				b.SyntaxDef = syndef
+				syntaxFile = f.Name()
+				break
+			} else {
+				syntaxFiles = append(syntaxFiles, syntaxFileBuffer{header, f.Name(), syndef})
+			}
 		}
 	}
 
-	// search in the default syntax files
-	for _, f := range config.ListRuntimeFiles(config.RTSyntaxHeader) {
-		data, err := f.Data()
-		if err != nil {
-			screen.TermMessage("Error loading syntax header file " + f.Name() + ": " + err.Error())
-			continue
-		}
+	if !foundDef {
+		// search in the default syntax files
+		for _, f := range config.ListRuntimeFiles(config.RTSyntaxHeader) {
+			data, err := f.Data()
+			if err != nil {
+				screen.TermMessage("Error loading syntax header file " + f.Name() + ": " + err.Error())
+				continue
+			}
 
-		header, err = highlight.MakeHeader(data)
-		if err != nil {
-			screen.TermMessage("Error reading syntax header file", f.Name(), err)
-			continue
-		}
+			header, err = highlight.MakeHeader(data)
+			if err != nil {
+				screen.TermMessage("Error reading syntax header file", f.Name(), err)
+				continue
+			}
 
-		if ft == "unknown" || ft == "" {
-			if highlight.MatchFiletype(header.FtDetect, b.Path, b.lines[0].data) {
+			if ft == "unknown" || ft == "" {
+				if header.MatchFileName(b.Path) {
+					syntaxFiles = append(syntaxFiles, syntaxFileBuffer{header, f.Name(), nil})
+				}
+			} else if header.FileType == ft {
 				syntaxFile = f.Name()
 				break
 			}
-		} else if header.FileType == ft {
-			syntaxFile = f.Name()
-			break
+		}
+	}
+
+	if syntaxFile == "" {
+		length := len(syntaxFiles)
+		if length > 0 {
+			signatureMatch := false
+			if length > 1 {
+				detectlimit := util.IntOpt(b.Settings["detectlimit"])
+				lineCount := len(b.lines)
+				limit := lineCount
+				if detectlimit > 0 && lineCount > detectlimit {
+					limit = detectlimit
+				}
+				for i := 0; i < length && !signatureMatch; i++ {
+					if syntaxFiles[i].header.HasFileSignature() {
+						for j := 0; j < limit && !signatureMatch; j++ {
+							if syntaxFiles[i].header.MatchFileSignature(b.lines[j].data) {
+								syntaxFile = syntaxFiles[i].fileName
+								b.SyntaxDef = syntaxFiles[i].syntaxDef
+								header = syntaxFiles[i].header
+								signatureMatch = true
+							}
+						}
+					}
+				}
+			}
+			if length == 1 || !signatureMatch {
+				syntaxFile = syntaxFiles[0].fileName
+				b.SyntaxDef = syntaxFiles[0].syntaxDef
+				header = syntaxFiles[0].header
+			}
 		}
 	}
 

internal/config/settings.go

@@ -45,6 +45,7 @@ func init() {
 var optionValidators = map[string]optionValidator{
 	"autosave":        validateNonNegativeValue,
 	"clipboard":       validateClipboard,
+	"detectlimit":     validateNonNegativeValue,
 	"tabsize":         validatePositiveValue,
 	"scrollmargin":    validateNonNegativeValue,
 	"scrollspeed":     validateNonNegativeValue,
@@ -282,6 +283,7 @@ var defaultCommonSettings = map[string]interface{}{
 	"basename":        false,
 	"colorcolumn":     float64(0),
 	"cursorline":      true,
+	"detectlimit":     float64(100),
 	"diffgutter":      false,
 	"encoding":        "utf-8",
 	"eofnewline":      true,

pkg/highlight/parser.go

@@ -33,27 +33,26 @@ func (g Group) String() string {
 // Then it has the rules which define how to highlight the file
 type Def struct {
 	*Header
-
 	rules *rules
 }
 
 type Header struct {
-	FileType string
-	FtDetect [2]*regexp.Regexp
+	FileType       string
+	FileNameRegex  *regexp.Regexp
+	SignatureRegex *regexp.Regexp
 }
 
 type HeaderYaml struct {
 	FileType string `yaml:"filetype"`
 	Detect   struct {
-		FNameRgx  string `yaml:"filename"`
-		HeaderRgx string `yaml:"header"`
+		FNameRegexStr     string `yaml:"filename"`
+		SignatureRegexStr string `yaml:"signature"`
 	} `yaml:"detect"`
 }
 
 type File struct {
 	FileType string
-
-	yamlSrc map[interface{}]interface{}
+	yamlSrc  map[interface{}]interface{}
 }
 
 // A Pattern is one simple syntax rule
@@ -103,14 +102,14 @@ func MakeHeader(data []byte) (*Header, error) {
 	header := new(Header)
 	var err error
 	header.FileType = string(lines[0])
-	fnameRgx := string(lines[1])
-	headerRgx := string(lines[2])
+	fnameRegexStr := string(lines[1])
+	signatureRegexStr := string(lines[2])
 
-	if fnameRgx != "" {
-		header.FtDetect[0], err = regexp.Compile(fnameRgx)
+	if fnameRegexStr != "" {
+		header.FileNameRegex, err = regexp.Compile(fnameRegexStr)
 	}
-	if err == nil && headerRgx != "" {
-		header.FtDetect[1], err = regexp.Compile(headerRgx)
+	if err == nil && signatureRegexStr != "" {
+		header.SignatureRegex, err = regexp.Compile(signatureRegexStr)
 	}
 
 	if err != nil {
@@ -132,11 +131,11 @@ func MakeHeaderYaml(data []byte) (*Header, error) {
 	header := new(Header)
 	header.FileType = hdrYaml.FileType
 
-	if hdrYaml.Detect.FNameRgx != "" {
-		header.FtDetect[0], err = regexp.Compile(hdrYaml.Detect.FNameRgx)
+	if hdrYaml.Detect.FNameRegexStr != "" {
+		header.FileNameRegex, err = regexp.Compile(hdrYaml.Detect.FNameRegexStr)
 	}
-	if err == nil && hdrYaml.Detect.HeaderRgx != "" {
-		header.FtDetect[1], err = regexp.Compile(hdrYaml.Detect.HeaderRgx)
+	if err == nil && hdrYaml.Detect.SignatureRegexStr != "" {
+		header.SignatureRegex, err = regexp.Compile(hdrYaml.Detect.SignatureRegexStr)
 	}
 
 	if err != nil {
@@ -146,6 +145,29 @@ func MakeHeaderYaml(data []byte) (*Header, error) {
 	return header, nil
 }
 
+// MatchFileName will check the given file name with the stored regex
+func (header *Header) MatchFileName(filename string) bool {
+	if header.FileNameRegex != nil {
+		return header.FileNameRegex.MatchString(filename)
+	}
+
+	return false
+}
+
+// HasFileSignature checks the presence of a stored signature
+func (header *Header) HasFileSignature() bool {
+	return header.SignatureRegex != nil
+}
+
+// MatchFileSignature will check the given line with the stored regex
+func (header *Header) MatchFileSignature(line []byte) bool {
+	if header.SignatureRegex != nil {
+		return header.SignatureRegex.Match(line)
+	}
+
+	return false
+}
+
 func ParseFile(input []byte) (f *File, err error) {
 	// This is just so if we have an error, we can exit cleanly and return the parse error to the user
 	defer func() {

runtime/help/colors.md

@@ -268,13 +268,13 @@ detect:
 ```
 
 Micro will match this regex against a given filename to detect the filetype.
-You may also provide an optional `header` regex that will check the first line
-of the file. For example:
+You may also provide an optional `signature` regex that will check a certain
+amount of lines of a file to find specific marks. For example:
 
 ```
 detect:
     filename: "\\.ya?ml$"
-    header: "%YAML"
+    signature: "%YAML"
 ```
 
 ### Syntax rules

runtime/help/options.md

@@ -100,6 +100,13 @@ Here are the available options:
 
 	default value: `true`
 
+* `detectlimit`: if this is not set to 0, it will limit the amount of first
+   lines in a file that are matched to determine the filetype.
+   A higher limit means better accuracy of guessing the filetype, but also
+   taking more time.
+
+   default value: `100`
+
 * `diffgutter`: display diff indicators before lines.
 
 	default value: `false`

runtime/syntax/PowerShell.yaml

@@ -5,7 +5,7 @@ filetype: powershell
 
 detect:
     filename: "\\.ps(1|m1|d1)$"
-    #header: ""
+    #signature: ""
 
 rules:
     # - comment.block:           # Block Comment

runtime/syntax/README.md

@@ -2,7 +2,7 @@
 
 Here are micro's syntax files.
 
-Each yaml file specifies how to detect the filetype based on file extension or headers (first line of the file).
+Each yaml file specifies how to detect the filetype based on file extension or given signature. The signature can be matched to all available lines of the file or to the value defined with the option `detectlimit` (to limit parse times) for a best "guess".
 Then there are patterns and regions linked to highlight groups which tell micro how to highlight that filetype.
 
 Making your own syntax files is very simple. I recommend you check the file after you are finished with the

runtime/syntax/awk.yaml

@@ -2,7 +2,7 @@ filetype: awk
 
 detect:
     filename: "\\.awk$"
-    header: "^#!.*bin/(env +)?awk( |$)"
+    signature: "^#!.*bin/(env +)?awk( |$)"
 
 rules:
     - preproc: "\\$[A-Za-z0-9_!@#$*?\\-]+"

runtime/syntax/bat.yaml

@@ -2,7 +2,7 @@ filetype: batch
 
 detect:
   filename: "(\\.bat$|\\.cmd$)"
-  # header: ""
+  # signature: ""
 
 rules:
   # Numbers

runtime/syntax/cpp.yaml

@@ -1,7 +1,8 @@
 filetype: c++
 
 detect:
-    filename: "(\\.c(c|pp|xx)$|\\.h(h|pp|xx)$|\\.ii?$|\\.(def)$)"
+    filename: "(\\.c(c|pp|xx)$|\\.h(h|pp|xx)?$|\\.ii?$|\\.(def)$)"
+    signature: "namespace|template|public|protected|private"
 
 rules:
     - identifier: "\\b[A-Z_][0-9A-Z_]*\\b"

runtime/syntax/crontab.yaml

@@ -2,7 +2,7 @@ filetype: crontab
 
 detect:
     filename: "crontab$"
-    header: "^#.*?/etc/crontab"
+    signature: "^#.*?/etc/crontab"
 
 rules:
       #              The time and date fields are:

runtime/syntax/csx.yaml

@@ -1,7 +1,7 @@
 filetype: csharp-script
 detect:
     filename: "\\.csx$"
-    header: "^#!.*/(env +)?dotnet-script( |$)"
+    signature: "^#!.*/(env +)?dotnet-script( |$)"
 
 rules:
     - include: "csharp"

runtime/syntax/fish.yaml

@@ -2,7 +2,7 @@ filetype: fish
 
 detect:
     filename: "\\.fish$"
-    header: "^#!.*/(env +)?fish( |$)"
+    signature: "^#!.*/(env +)?fish( |$)"
 
 rules:
       # Numbers

runtime/syntax/godoc.yaml

@@ -5,7 +5,7 @@ filetype: godoc
 
 detect:
     filename: "\\.godoc$"
-    header: package.*import
+    signature: package.*import
 
 rules:
     - preproc: "^[^ ].*"

runtime/syntax/groovy.yaml

@@ -2,7 +2,7 @@ filetype: groovy
 
 detect:
     filename: "(\\.(groovy|gy|gvy|gsh|gradle)$|^[Jj]enkinsfile$)"
-    header: "^#!.*/(env +)?groovy *$"
+    signature: "^#!.*/(env +)?groovy *$"
 
 rules:
     # And the style guide for constants is CONSTANT_CASE

runtime/syntax/html4.yaml

@@ -2,7 +2,7 @@ filetype: html4
 
 detect:
     filename: "\\.htm[l]?4$"
-    header: "<!DOCTYPE HTML PUBLIC \"-//W3C//DTD HTML 4.01//EN|http://www.w3.org/TR/html4/strict.dtd\">"
+    signature: "<!DOCTYPE HTML PUBLIC \"-//W3C//DTD HTML 4.01//EN|http://www.w3.org/TR/html4/strict.dtd\">"
 
 rules:
     - error: "<[^!].*?>"

runtime/syntax/html5.yaml

@@ -2,7 +2,7 @@ filetype: html5
 
 detect:
     filename: "\\.htm[l]?5$"
-    header: "<!DOCTYPE html5>"
+    signature: "<!DOCTYPE html5>"
 
 rules:
     - error: "<[^!].*?>"