#38 iteration 1: file attachments, 1 file per paste

- internal/store/blob.go: BlobStore interface + fs implementation with traversal-safe keys (<paste-id>/<sha256>), put/get/stat/delete - attachments table migration (id, paste_id, filename sanitized to 255, mime sniffed server-side, size, sha256, created_at) - POST /api/pastes now accepts multipart/form-data with a 'file' part; 1 file = 1 paste: file replaces text content when both are sent - 25 MB per-file limit enforced server-side (413 file_too_large) - GET /f/{attachment-id}/{filename}: stored sniffed mime, nosniff, inline only for images/pdf, html/svg/xml forced to text/plain (#34 rule) - paste view renders attachment chip + inline image preview - /new: dropzone with file picker, drag-and-drop, Ctrl+V file paste, file chip with name/size/remove, matches pill/radius design - tests: blob roundtrip/traversal/sanitize; multipart create (mime sniffing, client mime ignored, size limit, two-file reject, html/svg forcing, 404s, password/expiry fields)
2026-09-09 22:21:15 -05:00
parent 821f49993f
commit 4948ef9f1c
12 changed files with 1244 additions and 3 deletions
@@ -0,0 +1,351 @@
+package api
+
+import (
+	"bytes"
+	"fmt"
+	"io"
+	"net/http"
+	"strconv"
+	"strings"
+	"time"
+
+	"github.com/go-chi/chi/v5"
+
+	"palette/internal/store"
+)
+
+// #38: file attachments, iteration 1: one file per paste. A paste either has
+// text content OR one attached file. Multipart create + /f/ serving route.
+
+const (
+	MaxAttachmentBytes = 25 << 20                   // 25 MB per file
+	maxFileBytesHard   = MaxAttachmentBytes + 1<<20 // sniff headroom; over this reject before reading it all
+)
+
+// sniffMime runs http.DetectContentType on the first 512 bytes (and any
+// remainder of the head) of r, returning the sniffed mime and a reader that
+// replays the full stream. Mime is NEVER taken from the client.
+func sniffMime(r io.Reader) (string, io.Reader, error) {
+	head := make([]byte, 512)
+	n, err := io.ReadFull(r, head)
+	if err != nil && err != io.ErrUnexpectedEOF && err != io.EOF {
+		return "", nil, err
+	}
+	head = head[:n]
+	mime := http.DetectContentType(head)
+	return mime, io.MultiReader(bytes.NewReader(head), r), nil
+}
+
+// sanitizeMimeForServing maps the stored (sniffed) mime to the Content-Type
+// used on /f/. Active-content types (html, svg, xml...) are forced to
+// text/plain — same rule as the /raw #34 fix — so a malicious upload can
+// never execute on this origin.
+func serveContentType(mime string) string {
+	base := mime
+	if i := strings.IndexByte(mime, ';'); i >= 0 {
+		base = strings.TrimSpace(mime[:i])
+	}
+	base = strings.ToLower(base)
+	switch base {
+	case "text/html", "image/svg+xml", "application/xhtml+xml", "text/xml",
+		"application/xml", "application/xhtml", "image/xml+svg":
+		return "text/plain; charset=utf-8"
+	}
+	return mime
+}
+
+// inlineable reports whether the sniffed mime is safe to render inline
+// (Content-Disposition: inline); everything else downloads as an attachment.
+func inlineable(mime string) bool {
+	base := mime
+	if i := strings.IndexByte(mime, ';'); i >= 0 {
+		base = strings.TrimSpace(mime[:i])
+	}
+	base = strings.ToLower(base)
+	switch {
+	case strings.HasPrefix(base, "image/"), base == "application/pdf":
+		return true
+	}
+	return false
+}
+
+// mime/multipart parts are fully read during parsing (the multipart reader
+// closes each part when advancing) and the mime is sniffed from bytes.
+
+// limitAttachment rejects reads past the 25 MB per-file cap server-side.
+type limitReader struct {
+	r   io.Reader
+	n   int64
+	max int64
+}
+
+func (l *limitReader) Read(p []byte) (int, error) {
+	if l.n > l.max {
+		return 0, store.ErrFileTooLarge
+	}
+	n, err := l.r.Read(p)
+	l.n += int64(n)
+	if l.n > l.max && err == nil {
+		err = store.ErrFileTooLarge
+	}
+	return n, err
+}
+
+// handleCreatePasteMultipart implements POST /api/pastes with
+// multipart/form-data (#38). Fields mirror the JSON create path; a 'file'
+// part makes the paste a file paste (1 file = 1 paste: if text content is
+// also present, the file wins and the text is ignored — simplest correct
+// behavior, documented in the PR).
+func (a *apiServer) handleCreatePasteMultipart(w http.ResponseWriter, r *http.Request, s Settings) {
+	blobs := a.store.Blobs()
+	if blobs == nil {
+		writeErr(w, 500, "blob storage unavailable")
+		return
+	}
+	// guard the raw body: 25 MB file + multipart overhead headroom
+	mr, err := r.MultipartReader()
+	if err != nil {
+		writeErr(w, 400, "invalid multipart body")
+		return
+	}
+	var (
+		p        store.Paste
+		fileSeen bool
+		att      store.Attachment
+		fileBody io.Reader
+	)
+	for {
+		part, err := mr.NextPart()
+		if err == io.EOF {
+			break
+		}
+		if err != nil {
+			if isBodyTooLarge(err) {
+				writeErrCode(w, http.StatusRequestEntityTooLarge, "content_too_large", "request body too large")
+				return
+			}
+			writeErr(w, 400, "invalid multipart body")
+			return
+		}
+		name := part.FormName()
+		if name == "file" {
+			if fileSeen {
+				writeErrCode(w, 400, "one_file_only", "Choose either text or a file for now. Only one file per paste.")
+				part.Close()
+				return
+			}
+			// The part must be fully read during parsing: the multipart
+			// reader closes it as soon as the next part is fetched. Read it
+			// here into memory (bounded by the 25 MB cap) and sniff the mime
+			// from the content, never from client headers.
+			limited := io.LimitReader(part, MaxAttachmentBytes+1)
+			raw, err := io.ReadAll(limited)
+			part.Close()
+			if err != nil {
+				writeErr(w, 400, "invalid file part")
+				return
+			}
+			if int64(len(raw)) > MaxAttachmentBytes {
+				writeErrCode(w, http.StatusRequestEntityTooLarge, "file_too_large",
+					"File is too large. The limit is 25 MB.")
+				return
+			}
+			if len(raw) == 0 {
+				writeErrCode(w, 400, "content_empty", "The file is empty.")
+				return
+			}
+			filename := store.SanitizeFilename(part.FileName())
+			mime := http.DetectContentType(raw[:min(512, len(raw))])
+			att = store.Attachment{PasteID: "pending", Filename: filename, Mime: mime}
+			fileBody = bytes.NewReader(raw)
+			fileSeen = true
+			continue
+		}
+		val, err := io.ReadAll(io.LimitReader(part, 1<<16))
+		part.Close()
+		if err != nil {
+			writeErr(w, 400, "invalid multipart field")
+			return
+		}
+		v := string(val)
+		switch name {
+		case "content":
+			p.Content = v
+		case "title":
+			p.Title = &v
+		case "language":
+			p.Language = &v
+		case "custom_slug":
+			p.CustomSlug = &v
+		case "password":
+			p.Password = &v
+		case "expires_in":
+			p.ExpiresIn = &v
+		case "visibility":
+			p.Visibility = v
+		case "burn_after_read":
+			p.BurnAfterRead = v == "true" || v == "1" || v == "on"
+		case "burn_after_reads":
+			if n, err := strconv.Atoi(v); err == nil {
+				p.BurnAfterReads = &n
+			}
+		}
+	}
+
+	if fileSeen {
+		// 1 file = 1 paste: the file replaces text content.
+		p.Content = ""
+	} else if status, msg := checkContent(p.Content, s.MaxContentBytes); status != 0 {
+		if status == http.StatusRequestEntityTooLarge {
+			writeErrCode(w, status, "content_too_large", msg)
+		} else {
+			writeErrCode(w, status, "content_empty", msg)
+		}
+		return
+	}
+	// #86 metadata bounds + default expiry: same rules as the JSON path
+	if p.Title != nil {
+		t, err := checkTitle(*p.Title)
+		if err != nil {
+			writeErr(w, 400, err.Error())
+			return
+		}
+		p.Title = &t
+	}
+	if p.Language != nil {
+		l, err := checkLanguage(*p.Language)
+		if err != nil {
+			writeErr(w, 400, err.Error())
+			return
+		}
+		if l == "" {
+			p.Language = nil
+		} else {
+			p.Language = &l
+		}
+	}
+	if p.BurnAfterReads != nil {
+		if err := parseBurnAfterReads(*p.BurnAfterReads); err != nil {
+			writeErr(w, 400, err.Error())
+			return
+		}
+	}
+	if (p.ExpiresIn == nil || *p.ExpiresIn == "") && s.DefaultExpiry != "" {
+		def := s.DefaultExpiry
+		p.ExpiresIn = &def
+	}
+	p.ViewerID = currentViewerID(r)
+
+	created, err := a.store.CreatePaste(&p)
+	if err != nil {
+		writeErrCode(w, 400, createErrCode(err), err.Error())
+		return
+	}
+
+	resp := map[string]any{
+		"id":             created.ID,
+		"deletion_token": created.DeletionToken,
+		"url":            "/" + created.ID,
+		"raw_url":        "/raw/" + created.ID,
+		"api_url":        "/api/pastes/" + created.ID,
+		"expires_at":     created.ExpiresAt,
+		"created_at":     created.CreatedAt,
+		"rate_limit":     map[string]int{"create_per_sec": 1, "burst": 5},
+	}
+
+	if fileSeen {
+		att.PasteID = created.ID
+		// size pre-check happens inside the limited read; re-run with limit
+		// enforced so oversized uploads fail before the blob is stored.
+		err := a.store.CreateAttachment(&att, fileBody, blobs)
+		if err == store.ErrFileTooLarge {
+			a.store.SoftDelete(created.ID)
+			writeErrCode(w, http.StatusRequestEntityTooLarge, "file_too_large",
+				"File is too large. The limit is 25 MB.")
+			return
+		}
+		if err != nil {
+			a.store.SoftDelete(created.ID)
+			writeErr(w, 500, "could not store file")
+			return
+		}
+		resp["attachment"] = map[string]any{
+			"id":       att.ID,
+			"filename": att.Filename,
+			"mime":     att.Mime,
+			"size":     att.Size,
+			"sha256":   att.SHA256,
+			"url":      "/f/" + att.ID + "/" + att.Filename,
+		}
+	}
+	writeJSON(w, 201, resp)
+}
+
+// handleServeAttachment serves GET /f/{attachment-id}/{filename} with the
+// stored (server-sniffed) mime, nosniff, and a safe Content-Disposition.
+// The filename path segment is decorative; lookups key on the attachment id.
+func (a *apiServer) handleServeAttachment(w http.ResponseWriter, r *http.Request) {
+	id := chi.URLParam(r, "aid")
+	att, err := a.store.GetAttachment(id)
+	if err != nil {
+		writeErr(w, 500, "db error")
+		return
+	}
+	if att == nil {
+		writeErr(w, 404, "attachment not found")
+		return
+	}
+	// attachment inherits the paste's lifecycle: gone if the paste is gone
+	row, err := a.store.GetPaste(att.PasteID)
+	if err != nil || row == nil {
+		writeErr(w, 404, "attachment not found")
+		return
+	}
+	if row.ExpiresAt.Valid && row.ExpiresAt.Int64 < time.Now().Unix() {
+		writeErr(w, 404, "attachment not found")
+		return
+	}
+	if row.Burned() {
+		writeErr(w, 404, "attachment not found")
+		return
+	}
+
+	blobs := a.store.Blobs()
+	if blobs == nil {
+		writeErr(w, 500, "blob storage unavailable")
+		return
+	}
+	blob, err := blobs.Get(att.PasteID + "/" + att.SHA256)
+	if err != nil {
+		writeErr(w, 404, "attachment not found")
+		return
+	}
+	defer blob.Close()
+
+	ct := serveContentType(att.Mime)
+	w.Header().Set("Content-Type", ct)
+	w.Header().Set("X-Content-Type-Options", "nosniff")
+	disposition := "attachment"
+	if inlineable(att.Mime) {
+		disposition = "inline"
+	}
+	w.Header().Set("Content-Disposition",
+		fmt.Sprintf(`%s; filename="%s"`, disposition, asciiFilename(att.Filename)))
+	w.Header().Set("Content-Length", fmt.Sprintf("%d", att.Size))
+	http.ServeContent(w, r, "", time.Unix(att.CreatedAt, 0), blob)
+}
+
+// asciiFilename quotes a filename for the Content-Disposition header,
+// escaping quotes and backslashes and dropping non-ASCII bytes.
+func asciiFilename(name string) string {
+	var b strings.Builder
+	for _, r := range name {
+		if r < 128 && r != '"' && r != '\\' && r > 31 {
+			b.WriteRune(r)
+		}
+	}
+	if b.Len() == 0 {
+		return "file"
+	}
+	return b.String()
+}