From f6eedd4dc8fb10df869750c69e2bead0521ec0eb Mon Sep 17 00:00:00 2001 From: zeripath Date: Fri, 26 Apr 2019 13:00:30 +0100 Subject: [PATCH] UI: Detect and restore encoding and BOM in content (#6727) * detect and remove a decoded BOM Signed-off-by: Andrew Thornton * Restore the previous encoding and BOM * On error keep as UTF-8 Signed-off-by: Andrew Thornton * create remove BOM function * Deal with LFSed content * Update modules/repofiles/update.go * Fix final LFS bug * Keep LFS sections referring to opts.Content --- modules/base/tool.go | 12 +++++ modules/repofiles/update.go | 95 ++++++++++++++++++++++++++++++++++++- modules/templates/helper.go | 14 +++--- 3 files changed, 114 insertions(+), 7 deletions(-) diff --git a/modules/base/tool.go b/modules/base/tool.go index 97fd87e85c40..d99ed2bab4f1 100644 --- a/modules/base/tool.go +++ b/modules/base/tool.go @@ -5,6 +5,7 @@ package base import ( + "bytes" "crypto/md5" "crypto/rand" "crypto/sha1" @@ -36,6 +37,9 @@ import ( "github.com/gogits/chardet" ) +// UTF8BOM is the utf-8 byte-order marker +var UTF8BOM = []byte{'\xef', '\xbb', '\xbf'} + // EncodeMD5 encodes string to md5 hex value. func EncodeMD5(str string) string { m := md5.New() @@ -91,6 +95,14 @@ func DetectEncoding(content []byte) (string, error) { return result.Charset, err } +// RemoveBOMIfPresent removes a UTF-8 BOM from a []byte +func RemoveBOMIfPresent(content []byte) []byte { + if len(content) > 2 && bytes.Equal(content[0:3], UTF8BOM) { + return content[3:] + } + return content +} + // BasicAuthDecode decode basic auth string func BasicAuthDecode(encoded string) (string, string, error) { s, err := base64.StdEncoding.DecodeString(encoded) diff --git a/modules/repofiles/update.go b/modules/repofiles/update.go index e9b307753560..8bc3b50ae046 100644 --- a/modules/repofiles/update.go +++ b/modules/repofiles/update.go @@ -5,13 +5,19 @@ package repofiles import ( + "bytes" "fmt" "path" "strings" + "golang.org/x/net/html/charset" + "golang.org/x/text/transform" + "code.gitea.io/gitea/models" + "code.gitea.io/gitea/modules/base" "code.gitea.io/gitea/modules/git" "code.gitea.io/gitea/modules/lfs" + "code.gitea.io/gitea/modules/log" "code.gitea.io/gitea/modules/setting" "code.gitea.io/sdk/gitea" ) @@ -37,6 +43,70 @@ type UpdateRepoFileOptions struct { Committer *IdentityOptions } +func detectEncodingAndBOM(entry *git.TreeEntry, repo *models.Repository) (string, bool) { + reader, err := entry.Blob().DataAsync() + if err != nil { + // return default + return "UTF-8", false + } + defer reader.Close() + buf := make([]byte, 1024) + n, err := reader.Read(buf) + if err != nil { + // return default + return "UTF-8", false + } + buf = buf[:n] + + if setting.LFS.StartServer { + meta := lfs.IsPointerFile(&buf) + if meta != nil { + meta, err = repo.GetLFSMetaObjectByOid(meta.Oid) + if err != nil && err != models.ErrLFSObjectNotExist { + // return default + return "UTF-8", false + } + } + if meta != nil { + dataRc, err := lfs.ReadMetaObject(meta) + if err != nil { + // return default + return "UTF-8", false + } + defer dataRc.Close() + buf = make([]byte, 1024) + n, err = dataRc.Read(buf) + if err != nil { + // return default + return "UTF-8", false + } + buf = buf[:n] + } + + } + + encoding, err := base.DetectEncoding(buf) + if err != nil { + // just default to utf-8 and no bom + return "UTF-8", false + } + if encoding == "UTF-8" { + return encoding, bytes.Equal(buf[0:3], base.UTF8BOM) + } + charsetEncoding, _ := charset.Lookup(encoding) + if charsetEncoding == nil { + return "UTF-8", false + } + + result, n, err := transform.String(charsetEncoding.NewDecoder(), string(buf)) + + if n > 2 { + return encoding, bytes.Equal([]byte(result)[0:3], base.UTF8BOM) + } + + return encoding, false +} + // CreateOrUpdateRepoFile adds or updates a file in the given repository func CreateOrUpdateRepoFile(repo *models.Repository, doer *models.User, opts *UpdateRepoFileOptions) (*gitea.FileResponse, error) { // If no branch name is set, assume master @@ -118,6 +188,9 @@ func CreateOrUpdateRepoFile(repo *models.Repository, doer *models.User, opts *Up opts.LastCommitID = commit.ID.String() } + encoding := "UTF-8" + bom := false + if !opts.IsNewFile { fromEntry, err := commit.GetTreeEntryByPath(fromTreePath) if err != nil { @@ -151,6 +224,7 @@ func CreateOrUpdateRepoFile(repo *models.Repository, doer *models.User, opts *Up // haven't been made. We throw an error if one wasn't provided. return nil, models.ErrSHAOrCommitIDNotProvided{} } + encoding, bom = detectEncodingAndBOM(fromEntry, repo) } // For the path where this file will be created/updated, we need to make @@ -235,9 +309,28 @@ func CreateOrUpdateRepoFile(repo *models.Repository, doer *models.User, opts *Up } content := opts.Content + if bom { + content = string(base.UTF8BOM) + content + } + if encoding != "UTF-8" { + charsetEncoding, _ := charset.Lookup(encoding) + if charsetEncoding != nil { + result, _, err := transform.String(charsetEncoding.NewEncoder(), string(content)) + if err != nil { + // Look if we can't encode back in to the original we should just stick with utf-8 + log.Error("Error re-encoding %s (%s) as %s - will stay as UTF-8: %v", opts.TreePath, opts.FromTreePath, encoding, err) + result = content + } + content = result + } else { + log.Error("Unknown encoding: %s", encoding) + } + } + // Reset the opts.Content to our adjusted content to ensure that LFS gets the correct content + opts.Content = content var lfsMetaObject *models.LFSMetaObject - if filename2attribute2info[treePath] != nil && filename2attribute2info[treePath]["filter"] == "lfs" { + if setting.LFS.StartServer && filename2attribute2info[treePath] != nil && filename2attribute2info[treePath]["filter"] == "lfs" { // OK so we are supposed to LFS this data! oid, err := models.GenerateLFSOid(strings.NewReader(opts.Content)) if err != nil { diff --git a/modules/templates/helper.go b/modules/templates/helper.go index d3eb8c48b841..b6c5cc594549 100644 --- a/modules/templates/helper.go +++ b/modules/templates/helper.go @@ -267,7 +267,7 @@ func ToUTF8WithErr(content []byte) (string, error) { if err != nil { return "", err } else if charsetLabel == "UTF-8" { - return string(content), nil + return string(base.RemoveBOMIfPresent(content)), nil } encoding, _ := charset.Lookup(charsetLabel) @@ -277,19 +277,21 @@ func ToUTF8WithErr(content []byte) (string, error) { // If there is an error, we concatenate the nicely decoded part and the // original left over. This way we won't lose data. - result, n, err := transform.String(encoding.NewDecoder(), string(content)) + result, n, err := transform.Bytes(encoding.NewDecoder(), content) if err != nil { - result = result + string(content[n:]) + result = append(result, content[n:]...) } - return result, err + result = base.RemoveBOMIfPresent(result) + + return string(result), err } // ToUTF8WithFallback detects the encoding of content and coverts to UTF-8 if possible func ToUTF8WithFallback(content []byte) []byte { charsetLabel, err := base.DetectEncoding(content) if err != nil || charsetLabel == "UTF-8" { - return content + return base.RemoveBOMIfPresent(content) } encoding, _ := charset.Lookup(charsetLabel) @@ -304,7 +306,7 @@ func ToUTF8WithFallback(content []byte) []byte { return append(result, content[n:]...) } - return result + return base.RemoveBOMIfPresent(result) } // ToUTF8 converts content to UTF8 encoding and ignore error