byre/internal/server/files.go

package server

import (
	"bytes"
	"errors"
	"fmt"
	"io"
	"io/fs"
	"net/http"
	"os"
	"path"
	"path/filepath"
	"slices"
	"strings"
	"unicode/utf8"

	"g.fc.run/ai/byre/internal/api"
	"g.fc.run/ai/byre/internal/git"
)

const (
	maxFileBytes  = 256 << 10
	maxDirEntries = 5000
)

func (s *Server) GetTaskDir(w http.ResponseWriter, r *http.Request, taskID string, params api.GetTaskDirParams) {
	t, _, ok := s.activeTask(w, r, taskID)
	if !ok {
		return
	}
	rel := ""
	if params.Path != nil {
		rel = *params.Path
	}
	name, ok := worktreeName(w, rel)
	if !ok {
		return
	}

	root, err := os.OpenRoot(t.WorktreePath)
	if err != nil {
		writeError(w, http.StatusInternalServerError, err)
		return
	}
	defer root.Close()

	f, err := root.Open(name)
	if err != nil {
		writePathError(w, err)
		return
	}
	defer f.Close()

	fi, err := f.Stat()
	if err != nil {
		writeError(w, http.StatusInternalServerError, err)
		return
	}
	if !fi.IsDir() {
		writeError(w, http.StatusBadRequest, fmt.Errorf("%s is not a directory", rel))
		return
	}

	des, err := f.ReadDir(-1)
	if err != nil {
		writeError(w, http.StatusInternalServerError, err)
		return
	}

	entries := []api.DirEntry{}
	for _, de := range des {
		if de.Name() == ".git" {
			continue
		}
		entries = append(entries, dirEntry(root, path.Join(name, de.Name()), de))
	}
	slices.SortFunc(entries, func(a, b api.DirEntry) int {
		if (a.Type == api.DirEntryTypeDir) != (b.Type == api.DirEntryTypeDir) {
			if a.Type == api.DirEntryTypeDir {
				return -1
			}
			return 1
		}
		return strings.Compare(a.Name, b.Name)
	})

	listing := api.DirListing{Entries: entries}
	if name != "." {
		listing.Path = name
	}
	if len(entries) > maxDirEntries {
		listing.Entries = entries[:maxDirEntries]
		listing.Truncated = true
	}

	real, err := realDir(t.WorktreePath, name)
	if err != nil {
		writeError(w, http.StatusInternalServerError, err)
		return
	}
	paths := []string{}
	for _, e := range listing.Entries {
		paths = append(paths, path.Join(real, e.Name))
	}
	ignored, err := git.Ignored(t.WorktreePath, paths)
	if err != nil {
		writeError(w, http.StatusInternalServerError, err)
		return
	}
	for i, p := range paths {
		listing.Entries[i].Ignored = ignored[p]
	}
	writeJSON(w, http.StatusOK, listing)
}

func realDir(worktree, name string) (string, error) {
	root, err := filepath.EvalSymlinks(worktree)
	if err != nil {
		return "", err
	}
	dir, err := filepath.EvalSymlinks(filepath.Join(worktree, name))
	if err != nil {
		return "", err
	}
	rel, err := filepath.Rel(root, dir)
	if err != nil {
		return "", err
	}
	return filepath.ToSlash(rel), nil
}

func dirEntry(root *os.Root, name string, de fs.DirEntry) api.DirEntry {
	e := api.DirEntry{Name: de.Name(), Type: api.DirEntryTypeOther}

	var fi fs.FileInfo
	var err error
	if de.Type()&fs.ModeSymlink != 0 {
		e.Symlink = true
		fi, err = root.Stat(name)
	} else {
		fi, err = de.Info()
	}
	if err != nil {
		return e
	}

	switch {
	case fi.IsDir():
		e.Type = api.DirEntryTypeDir
	case fi.Mode().IsRegular():
		e.Type = api.DirEntryTypeFile
		e.Size = fi.Size()
	}
	return e
}

func (s *Server) GetTaskFile(w http.ResponseWriter, r *http.Request, taskID string, params api.GetTaskFileParams) {
	t, _, ok := s.activeTask(w, r, taskID)
	if !ok {
		return
	}
	if params.Path == "" {
		writeError(w, http.StatusBadRequest, errors.New("path is required"))
		return
	}
	name, ok := worktreeName(w, params.Path)
	if !ok {
		return
	}

	root, err := os.OpenRoot(t.WorktreePath)
	if err != nil {
		writeError(w, http.StatusInternalServerError, err)
		return
	}
	defer root.Close()

	fi, err := root.Stat(name)
	if err != nil {
		writePathError(w, err)
		return
	}
	if !fi.Mode().IsRegular() {
		writeError(w, http.StatusBadRequest, fmt.Errorf("%s is not a regular file", params.Path))
		return
	}

	f, err := root.Open(name)
	if err != nil {
		writePathError(w, err)
		return
	}
	defer f.Close()

	b, err := io.ReadAll(io.LimitReader(f, maxFileBytes+1))
	if err != nil {
		writeError(w, http.StatusInternalServerError, err)
		return
	}

	fc := api.FileContent{Path: name, Size: fi.Size()}
	if len(b) > maxFileBytes {
		fc.Truncated = true
		b = b[:maxFileBytes]
		for n := 0; n < utf8.UTFMax-1 && len(b) > 0 && !utf8.Valid(b); n++ {
			b = b[:len(b)-1]
		}
	}
	if bytes.IndexByte(b, 0) >= 0 || !utf8.Valid(b) {
		fc.Binary = true
		writeJSON(w, http.StatusOK, fc)
		return
	}
	fc.Text = string(b)
	writeJSON(w, http.StatusOK, fc)
}

func worktreeName(w http.ResponseWriter, rel string) (string, bool) {
	if rel == "" {
		return ".", true
	}
	name := filepath.FromSlash(rel)
	if !filepath.IsLocal(name) {
		writeError(w, http.StatusBadRequest, fmt.Errorf("%s is not a path inside the worktree", rel))
		return "", false
	}
	name = filepath.ToSlash(filepath.Clean(name))
	if slices.Contains(strings.Split(name, "/"), ".git") {
		writeError(w, http.StatusBadRequest, fmt.Errorf("%s is inside .git", rel))
		return "", false
	}
	return name, true
}

func writePathError(w http.ResponseWriter, err error) {
	if errors.Is(err, fs.ErrNotExist) {
		writeError(w, http.StatusNotFound, err)
		return
	}
	writeError(w, http.StatusBadRequest, err)
}