gg/main.go

/*
gg is a recursive grep written in Go, with some shortcuts for everyday use.
*/
package main

import (
	"errors"
	"fmt"
	"os"
	"path/filepath"
	"regexp"
	"strings"

	"github.com/bmatcuk/doublestar/v4"
	"github.com/spf13/cobra"
	"golang.org/x/sys/unix"
)

// TODO:
//  - configurable defaults for exclude.

func main() {
	if err := rootCmd.Execute(); err != nil {
		os.Exit(1)
	}
}

var rootCmd = &cobra.Command{
	Use:   "gg pattern [path1 [path2 …]]",
	Short: "gg is a recursive grep",
	Long: `gg is a recursive grep. Given a regexp (or fixed pattern) it will search for
the pattern recursively in the current working directory. It will print a
coloured header per file along with the matching line and pattern.

It is possible to scan specific files or directories, rather than the default
current working directory. To do this, simply specify the path(s) as arguments
following the pattern.

It is possible to scan for multiple patterns using the -e (or -Q) argument,
which can be repeated multiple times. -e specifies a regular expression and
-Q a fixed pattern. When using either flag, any non-flag arguments are treated
as paths to scan.

Search defaults to case-sensitive but the -i flag may be passed to make all
search terms case-insensitive. Alternatively, the "(?i)" construct may be added
to a regular expression to make that specific expression case insensitive.

Files and directories can be excluded with the -x option. This supports bash-style
globs with '*', '?', '[a-z]', '{this,that}', or '/**/' to match zero or more
directories. By default, .git and vim swap files are ignored. Similarly, -I
filters files to include. Examples:

	# ignore files/dirs with .js or .css suffix
	gg -x '*.js' -x '*.css' pattern

	# only match files with .go suffix (any subdir)
	gg -I '*.go' pattern

	# only match files whose parent dir is "stuff", but ignore "foo" subdir
	gg -x ./foo -I 'stuff/*' pattern

	# only match .js files with a directory "things" in the path, but ignore
	# .min.js (e.g. will match "foo/things/bar/my.js")
	gg -I 'things/**/*.js' -x '*.min.js' pattern

Symlinks named on the command line are followed, but by default symlinks are
not followed when recursing into directories. -L allows them to be
dereferenced.`,

	RunE: run,
}

var (
	// flags
	searchRegexp   []string
	searchFixed    []string
	searchPath     []string
	excludeList    []string
	includeList    []string
	ignoreCase     bool
	noColour       bool
	binaryFile     notPlainTextFlag
	minifiedFile   notPlainTextFlag
	followSymlinks bool

	// computed from searchRegexp, searchFixed. Each regexp here will be
	// matched against each line of each input file.
	regexps []*regexp.Regexp

	// formats output
	display *Display
)

func init() {
	rootCmd.Flags().StringSliceVarP(&searchRegexp, "grep", "e", nil, "pattern to match (regular expression)")
	rootCmd.Flags().StringSliceVarP(&searchFixed, "fixed", "Q", nil, "pattern to match (fixed string)")
	rootCmd.Flags().StringSliceVarP(&excludeList, "exclude", "x", []string{".git", ".*.swp"}, "files/directories to exclude")
	rootCmd.Flags().StringSliceVarP(&includeList, "include", "I", nil, "files/directories to include")
	rootCmd.Flags().BoolVarP(&ignoreCase, "ignore-case", "i", false, "make all searches case insensitive")
	rootCmd.Flags().BoolVarP(&noColour, "no-colour", "C", false, "disable colour output")
	rootCmd.Flags().Var(&binaryFile, "binary", "what to do with binary files")
	rootCmd.Flags().Var(&minifiedFile, "minified", "what to do with minified text files")
	rootCmd.Flags().BoolVarP(&followSymlinks, "dereference", "L", false, "follow symlinks when recursing")
}

func run(c *cobra.Command, args []string) error {
	display = NewDisplay(noColour)

	// if no -e or -Q flag is passed, then the first arg is taken to be
	// the pattern to match
	if len(searchRegexp) == 0 && len(searchFixed) == 0 {
		if len(args) == 0 {
			return errors.New("no pattern specified")
		}
		searchRegexp = append(searchRegexp, args[0])
		args = args[1:]
	}

	// remaining arguments are treated as search paths; an empty list is
	// taken to mean the CWD
	searchPath = args
	if len(searchPath) == 0 {
		searchPath = append(searchPath, ".")
	}

	// if we got past argument passing, then returned errors are runtime
	// things (like file not found) that shouldn't trigger a usage message.
	c.SilenceUsage = true

	// for -x and -I, an undecorated pattern is treated as a suffix match
	for i, x := range excludeList {
		if !strings.HasPrefix(x, "**/") && !strings.HasPrefix(x, "./") {
			x = "**/" + x
			excludeList[i] = x
		}
		if !doublestar.ValidatePattern(x) {
			return fmt.Errorf("invalid exclude pattern %q", x)
		}
	}
	for i, x := range includeList {
		if !strings.HasPrefix(x, "**/") && !strings.HasPrefix(x, "./") {
			x = "**/" + x
			includeList[i] = x
		}
		if !doublestar.ValidatePattern(x) {
			return fmt.Errorf("invalid include pattern %q", x)
		}
	}

	// compile regular expressions for matching
	for _, r := range searchRegexp {
		if ignoreCase {
			r = "(?i)" + r
		}
		re, err := regexp.Compile(r)
		if err != nil {
			return err
		}
		regexps = append(regexps, re)
	}
	for _, r := range searchFixed {
		r = regexp.QuoteMeta(r)
		if ignoreCase {
			r = "(?i)" + r
		}
		re, err := regexp.Compile(r)
		if err != nil {
		}
		regexps = append(regexps, re)
	}

	// search over named paths
	var errs []error
	for _, path := range searchPath {
		if err := search(path, true); err != nil {
			errs = append(errs, err)
		}
	}

	return errors.Join(errs...)
}

func recurse(path string) error {
	d, err := os.ReadDir(path)
	if err != nil {
		return err
	}

	var errs []error
	for _, de := range d {
		fullPath := filepath.Join(path, de.Name())
		if !shouldSearch(fullPath, de.IsDir()) {
			continue
		}
		if err := search(fullPath, followSymlinks); err != nil {
			errs = append(errs, err)
		}
	}
	return errors.Join(errs...)
}

// shouldSearch matches the full path of the file against the include and
// exclude lists, returning true if we should consider the file/directory for
// searching and false if not.
func shouldSearch(fullPath string, isDir bool) bool {
	// process the exclude list first
	for _, x := range excludeList {
		if exclude, _ := doublestar.Match(x, fullPath); exclude {
			return false
		}
	}

	// if the include list is empty, everything is included
	if len(includeList) == 0 {
		return true
	}

	for _, x := range includeList {
		match, _ := doublestar.Match(x, fullPath)
		fmt.Printf("[DEBUG] x=%q fullPath=%q isDir=%t match=%t\n", x, fullPath, isDir, match)

		// if it's a directory, and we have at least one recursive
		// matcher, then search
		if isDir && strings.HasPrefix(x, "**/") {
			return true
		}

		if include, _ := doublestar.Match(x, fullPath); include {
			return true
		}
	}

	return false
}

func search(path string, deref bool) error {
	var (
		st  os.FileInfo
		err error
	)

	if deref {
		st, err = os.Stat(path)
	} else {
		st, err = os.Lstat(path)
	}
	if err != nil {
		return err
	}

	switch {
	case st.IsDir():
		return recurse(path)

	case !st.Mode().IsRegular(),
		st.Size() == 0:
		return nil
	}

	f, err := os.Open(path)
	if err != nil {
		return err
	}
	defer f.Close()
	data, err := unix.Mmap(int(f.Fd()), 0, int(st.Size()), unix.PROT_READ, unix.MAP_PRIVATE)
	if err != nil {
		return err
	}
	defer unix.Munmap(data)

	file(path, data)
	return nil
}