Read expected buffer size (#17409) (#17430)

Backport of #17409

* Read expected buffer size.

* Changed name.

diff --git a/modules/charset/charset.go b/modules/charset/charset.go
index 3000864c2ea0e75278a3674bd365477ee099424f..55e183ebfc7cef671ad82aa5dcdb53e192d1e432 100644 (file)
--- a/modules/charset/charset.go
+++ b/modules/charset/charset.go
@@ -14,6 +14,7 @@ import (
 
        "code.gitea.io/gitea/modules/log"
        "code.gitea.io/gitea/modules/setting"
+       "code.gitea.io/gitea/modules/util"
 
        "github.com/gogs/chardet"
        "golang.org/x/net/html/charset"
@@ -26,9 +27,9 @@ var UTF8BOM = []byte{'\xef', '\xbb', '\xbf'}
 // ToUTF8WithFallbackReader detects the encoding of content and coverts to UTF-8 reader if possible
 func ToUTF8WithFallbackReader(rd io.Reader) io.Reader {
        var buf = make([]byte, 2048)
-       n, err := rd.Read(buf)
+       n, err := util.ReadAtMost(rd, buf)
        if err != nil {
-               return rd
+               return io.MultiReader(bytes.NewReader(RemoveBOMIfPresent(buf[:n])), rd)
        }
 
        charsetLabel, err := DetectEncoding(buf[:n])

diff --git a/modules/csv/csv.go b/modules/csv/csv.go
index ee54452891923ed473c6d7735eeb2a6150fe84bd..83b1a5d37bc989c34795cf72727db6a300083af2 100644 (file)
--- a/modules/csv/csv.go
+++ b/modules/csv/csv.go
@@ -30,11 +30,8 @@ func CreateReader(input io.Reader, delimiter rune) *stdcsv.Reader {
 // CreateReaderAndGuessDelimiter tries to guess the field delimiter from the content and creates a csv.Reader.
 func CreateReaderAndGuessDelimiter(rd io.Reader) (*stdcsv.Reader, error) {
        var data = make([]byte, 1e4)
-       size, err := rd.Read(data)
+       size, err := util.ReadAtMost(rd, data)
        if err != nil {
-               if err == io.EOF {
-                       return CreateReader(bytes.NewReader([]byte{}), rune(',')), nil
-               }
                return nil, err
        }
 

diff --git a/modules/git/blob.go b/modules/git/blob.go
index 5831bc3735aaebe929e83d5c675b85fc5630b2d8..39b28cbe26224eaab31d4eeb532ba4a7c1226394 100644 (file)
--- a/modules/git/blob.go
+++ b/modules/git/blob.go
@@ -12,6 +12,7 @@ import (
        "io/ioutil"
 
        "code.gitea.io/gitea/modules/typesniffer"
+       "code.gitea.io/gitea/modules/util"
 )
 
 // This file contains common functions between the gogit and !gogit variants for git Blobs
@@ -29,7 +30,7 @@ func (b *Blob) GetBlobContent() (string, error) {
        }
        defer dataRc.Close()
        buf := make([]byte, 1024)
-       n, _ := dataRc.Read(buf)
+       n, _ := util.ReadAtMost(dataRc, buf)
        buf = buf[:n]
        return string(buf), nil
 }

diff --git a/modules/repofiles/update.go b/modules/repofiles/update.go
index ad984c465ae8675fccec42f1a0673860a91bbf80..5b45479f3f6c207f577d211ba886e6973bea4ddc 100644 (file)
--- a/modules/repofiles/update.go
+++ b/modules/repofiles/update.go
@@ -19,6 +19,7 @@ import (
        repo_module "code.gitea.io/gitea/modules/repository"
        "code.gitea.io/gitea/modules/setting"
        "code.gitea.io/gitea/modules/structs"
+       "code.gitea.io/gitea/modules/util"
 
        stdcharset "golang.org/x/net/html/charset"
        "golang.org/x/text/transform"
@@ -61,7 +62,7 @@ func detectEncodingAndBOM(entry *git.TreeEntry, repo *models.Repository) (string
        }
        defer reader.Close()
        buf := make([]byte, 1024)
-       n, err := reader.Read(buf)
+       n, err := util.ReadAtMost(reader, buf)
        if err != nil {
                // return default
                return "UTF-8", false
@@ -84,7 +85,7 @@ func detectEncodingAndBOM(entry *git.TreeEntry, repo *models.Repository) (string
                                }
                                defer dataRc.Close()
                                buf = make([]byte, 1024)
-                               n, err = dataRc.Read(buf)
+                               n, err = util.ReadAtMost(dataRc, buf)
                                if err != nil {
                                        // return default
                                        return "UTF-8", false

diff --git a/modules/typesniffer/typesniffer.go b/modules/typesniffer/typesniffer.go
index d257b8179b0ad2aa1fc94408c064c737a23666cc..9e29b3557c4ac6b4ab4da55f506dca7c2a4d73fc 100644 (file)
--- a/modules/typesniffer/typesniffer.go
+++ b/modules/typesniffer/typesniffer.go
@@ -10,6 +10,8 @@ import (
        "net/http"
        "regexp"
        "strings"
+
+       "code.gitea.io/gitea/modules/util"
 )
 
 // Use at most this many bytes to determine Content Type.
@@ -86,8 +88,8 @@ func DetectContentType(data []byte) SniffedType {
 // DetectContentTypeFromReader guesses the content type contained in the reader.
 func DetectContentTypeFromReader(r io.Reader) (SniffedType, error) {
        buf := make([]byte, sniffLen)
-       n, err := r.Read(buf)
-       if err != nil && err != io.EOF {
+       n, err := util.ReadAtMost(r, buf)
+       if err != nil {
                return SniffedType{}, fmt.Errorf("DetectContentTypeFromReader io error: %w", err)
        }
        buf = buf[:n]

diff --git a/modules/util/io.go b/modules/util/io.go
new file mode 100644 (file)
index 0000000..b467c0a
--- /dev/null
+++ b/modules/util/io.go
@@ -0,0 +1,20 @@
+// Copyright 2021 The Gitea Authors. All rights reserved.
+// Use of this source code is governed by a MIT-style
+// license that can be found in the LICENSE file.
+
+package util
+
+import (
+       "io"
+)
+
+// ReadAtMost reads at most len(buf) bytes from r into buf.
+// It returns the number of bytes copied. n is only less then len(buf) if r provides fewer bytes.
+// If EOF occurs while reading, err will be nil.
+func ReadAtMost(r io.Reader, buf []byte) (n int, err error) {
+       n, err = io.ReadFull(r, buf)
+       if err == io.EOF || err == io.ErrUnexpectedEOF {
+               err = nil
+       }
+       return
+}

diff --git a/routers/common/repo.go b/routers/common/repo.go
index 8d33fb07fbbd5f1252c2e7ed8833c1a36561e6ec..e5b1a0493a191a2cce6d1956b394a3dc0fb6c512 100644 (file)
--- a/routers/common/repo.go
+++ b/routers/common/repo.go
@@ -18,6 +18,7 @@ import (
        "code.gitea.io/gitea/modules/log"
        "code.gitea.io/gitea/modules/setting"
        "code.gitea.io/gitea/modules/typesniffer"
+       "code.gitea.io/gitea/modules/util"
 )
 
 // ServeBlob download a git.Blob
@@ -42,8 +43,8 @@ func ServeBlob(ctx *context.Context, blob *git.Blob) error {
 // ServeData download file from io.Reader
 func ServeData(ctx *context.Context, name string, size int64, reader io.Reader) error {
        buf := make([]byte, 1024)
-       n, err := reader.Read(buf)
-       if err != nil && err != io.EOF {
+       n, err := util.ReadAtMost(reader, buf)
+       if err != nil {
                return err
        }
        if n >= 0 {

diff --git a/routers/web/repo/attachment.go b/routers/web/repo/attachment.go
index 5becbea2713a1a9f0fb9e15eafefc83a2e18c1ce..a32bc3e140e77dbba1c27686445f4682280b32e1 100644 (file)
--- a/routers/web/repo/attachment.go
+++ b/routers/web/repo/attachment.go
@@ -15,6 +15,7 @@ import (
        "code.gitea.io/gitea/modules/setting"
        "code.gitea.io/gitea/modules/storage"
        "code.gitea.io/gitea/modules/upload"
+       "code.gitea.io/gitea/modules/util"
        "code.gitea.io/gitea/routers/common"
 )
 
@@ -43,10 +44,8 @@ func uploadAttachment(ctx *context.Context, allowedTypes string) {
        defer file.Close()
 
        buf := make([]byte, 1024)
-       n, _ := file.Read(buf)
-       if n > 0 {
-               buf = buf[:n]
-       }
+       n, _ := util.ReadAtMost(file, buf)
+       buf = buf[:n]
 
        err = upload.Verify(buf, header.Filename, allowedTypes)
        if err != nil {

diff --git a/routers/web/repo/editor.go b/routers/web/repo/editor.go
index 0f978c7b01c599a7792a513fc0ec04454d23ec69..f7a8778344fd608c7e68144e6fae7adcb7350fcd 100644 (file)
--- a/routers/web/repo/editor.go
+++ b/routers/web/repo/editor.go
@@ -114,7 +114,7 @@ func editFile(ctx *context.Context, isNewFile bool) {
                ctx.Data["FileName"] = blob.Name()
 
                buf := make([]byte, 1024)
-               n, _ := dataRc.Read(buf)
+               n, _ := util.ReadAtMost(dataRc, buf)
                buf = buf[:n]
 
                // Only some file types are editable online as text.
@@ -747,7 +747,7 @@ func UploadFileToServer(ctx *context.Context) {
        defer file.Close()
 
        buf := make([]byte, 1024)
-       n, _ := file.Read(buf)
+       n, _ := util.ReadAtMost(file, buf)
        if n > 0 {
                buf = buf[:n]
        }

diff --git a/routers/web/repo/lfs.go b/routers/web/repo/lfs.go
index 173ffb773f8809512ab8a075a4139bc4ee3e66e0..937d623e5a868251b7498685efbf7caae6f4bcc3 100644 (file)
--- a/routers/web/repo/lfs.go
+++ b/routers/web/repo/lfs.go
@@ -26,6 +26,7 @@ import (
        "code.gitea.io/gitea/modules/setting"
        "code.gitea.io/gitea/modules/storage"
        "code.gitea.io/gitea/modules/typesniffer"
+       "code.gitea.io/gitea/modules/util"
 )
 
 const (
@@ -272,7 +273,7 @@ func LFSFileGet(ctx *context.Context) {
        }
        defer dataRc.Close()
        buf := make([]byte, 1024)
-       n, err := dataRc.Read(buf)
+       n, err := util.ReadAtMost(dataRc, buf)
        if err != nil {
                ctx.ServerError("Data", err)
                return
@@ -297,10 +298,10 @@ func LFSFileGet(ctx *context.Context) {
                        break
                }
 
-               buf := charset.ToUTF8WithFallbackReader(io.MultiReader(bytes.NewReader(buf), dataRc))
+               rd := charset.ToUTF8WithFallbackReader(io.MultiReader(bytes.NewReader(buf), dataRc))
 
                // Building code view blocks with line number on server side.
-               fileContent, _ := ioutil.ReadAll(buf)
+               fileContent, _ := ioutil.ReadAll(rd)
 
                var output bytes.Buffer
                lines := strings.Split(string(fileContent), "\n")

diff --git a/routers/web/repo/view.go b/routers/web/repo/view.go
index 2be81d7485cbb5493d154527c0016c7e75989c55..64b97a91abe92fdd88b11da1bdc748681d7828c0 100644 (file)
--- a/routers/web/repo/view.go
+++ b/routers/web/repo/view.go
@@ -31,6 +31,7 @@ import (
        "code.gitea.io/gitea/modules/setting"
        "code.gitea.io/gitea/modules/structs"
        "code.gitea.io/gitea/modules/typesniffer"
+       "code.gitea.io/gitea/modules/util"
 )
 
 const (
@@ -264,7 +265,7 @@ func renderDirectory(ctx *context.Context, treeLink string) {
                defer dataRc.Close()
 
                buf := make([]byte, 1024)
-               n, _ := dataRc.Read(buf)
+               n, _ := util.ReadAtMost(dataRc, buf)
                buf = buf[:n]
 
                st := typesniffer.DetectContentType(buf)
@@ -299,7 +300,7 @@ func renderDirectory(ctx *context.Context, treeLink string) {
                                        defer dataRc.Close()
 
                                        buf = make([]byte, 1024)
-                                       n, err = dataRc.Read(buf)
+                                       n, err = util.ReadAtMost(dataRc, buf)
                                        if err != nil {
                                                ctx.ServerError("Data", err)
                                                return
@@ -413,7 +414,7 @@ func renderFile(ctx *context.Context, entry *git.TreeEntry, treeLink, rawLink st
        ctx.Data["RawFileLink"] = rawLink + "/" + ctx.Repo.TreePath
 
        buf := make([]byte, 1024)
-       n, _ := dataRc.Read(buf)
+       n, _ := util.ReadAtMost(dataRc, buf)
        buf = buf[:n]
 
        st := typesniffer.DetectContentType(buf)
@@ -445,10 +446,8 @@ func renderFile(ctx *context.Context, entry *git.TreeEntry, treeLink, rawLink st
                                defer dataRc.Close()
 
                                buf = make([]byte, 1024)
-                               n, err = dataRc.Read(buf)
-                               // Error EOF don't mean there is an error, it just means we read to
-                               // the end
-                               if err != nil && err != io.EOF {
+                               n, err = util.ReadAtMost(dataRc, buf)
+                               if err != nil {
                                        ctx.ServerError("Data", err)
                                        return
                                }