From 3424d3057f7a63ad131f6bbddbb3c1d818dc3bb6 Mon Sep 17 00:00:00 2001
From: Eric Coissac <eric@coissac.eu>
Date: Wed, 14 May 2025 14:53:25 +0200
Subject: [PATCH] Changes to be committed: 	modified:  
 pkg/obiformats/ngsfilter_read.go 	modified:   pkg/obioptions/version.go 
 modified:   pkg/obiutils/mimetypes.go

---
 pkg/obiformats/ngsfilter_read.go | 11 +++++++----
 pkg/obioptions/version.go        |  2 +-
 pkg/obiutils/mimetypes.go        | 29 +++++++++++++++++++++++++----
 3 files changed, 33 insertions(+), 9 deletions(-)

diff --git a/pkg/obiformats/ngsfilter_read.go b/pkg/obiformats/ngsfilter_read.go
index fb409ce..df73a71 100644
--- a/pkg/obiformats/ngsfilter_read.go
+++ b/pkg/obiformats/ngsfilter_read.go
@@ -89,7 +89,7 @@ func _parseMainNGSFilter(text string) (obingslibrary.PrimerPair, obingslibrary.T
 }
 
 func NGSFilterCsvDetector(raw []byte, limit uint32) bool {
-	r := csv.NewReader(bytes.NewReader(obiutils.DropLastLine(raw, limit)))
+	r := csv.NewReader(bytes.NewReader(obiutils.DropLastLine(raw)))
 	r.Comma = ','
 	r.ReuseRecord = true
 	r.LazyQuotes = true
@@ -110,7 +110,6 @@ func NGSFilterCsvDetector(raw []byte, limit uint32) bool {
 		if err != nil {
 			return false
 		}
-
 		if nfields == 0 {
 			nfields = len(rec)
 		} else if nfields != len(rec) {
@@ -133,16 +132,20 @@ func OBIMimeNGSFilterTypeGuesser(stream io.Reader) (*mimetype.MIME, io.Reader, e
 		return nil, nil, err
 	}
 
+	buf = buf[:n]
+
+	obiutils.HasBOM(buf)
+
 	mimetype.Lookup("text/plain").Extend(NGSFilterCsvDetector, "text/ngsfilter-csv", ".csv")
 
 	// Detect the MIME type using the mimetype library
-	mimeType := mimetype.Detect(buf[:n])
+	mimeType := mimetype.Detect(buf)
 	if mimeType == nil {
 		return nil, nil, err
 	}
 
 	// Create a new reader based on the read data
-	newReader := io.Reader(bytes.NewReader(buf[:n]))
+	newReader := io.Reader(bytes.NewReader(buf))
 
 	if err == nil {
 		newReader = io.MultiReader(newReader, stream)
diff --git a/pkg/obioptions/version.go b/pkg/obioptions/version.go
index 33cb754..92b90d4 100644
--- a/pkg/obioptions/version.go
+++ b/pkg/obioptions/version.go
@@ -8,7 +8,7 @@ import (
 // corresponds to the last commit, and not the one when the file will be
 // commited
 
-var _Commit = "f1b9ac4"
+var _Commit = "f9324dd"
 var _Version = "Release 4.4.0"
 
 // Version returns the version of the obitools package.
diff --git a/pkg/obiutils/mimetypes.go b/pkg/obiutils/mimetypes.go
index 3192764..4d53c49 100644
--- a/pkg/obiutils/mimetypes.go
+++ b/pkg/obiutils/mimetypes.go
@@ -8,12 +8,33 @@ import (
 	"regexp"
 
 	"github.com/gabriel-vasile/mimetype"
+	log "github.com/sirupsen/logrus"
 )
 
-func DropLastLine(b []byte, readLimit uint32) []byte {
-	if readLimit == 0 || uint32(len(b)) < readLimit {
-		return b
+func HasBOM(data []byte) bool {
+	switch {
+	case bytes.HasPrefix(data, []byte{0xEF, 0xBB, 0xBF}):
+		log.Infoln("BOM detected: UTF-8 (EF BB BF)")
+		return true
+	case bytes.HasPrefix(data, []byte{0xFE, 0xFF}):
+		log.Infoln("BOM detected: UTF-16 Big Endian (FE FF)")
+		return true
+	case bytes.HasPrefix(data, []byte{0xFF, 0xFE}):
+		log.Infoln("BOM detected: UTF-16 Little Endian (FF FE)")
+		return true
+	case bytes.HasPrefix(data, []byte{0x00, 0x00, 0xFE, 0xFF}):
+		log.Infoln("BOM detected: UTF-32 Big Endian (00 00 FE FF)")
+		return true
+	case bytes.HasPrefix(data, []byte{0xFF, 0xFE, 0x00, 0x00}):
+		log.Infoln("BOM detected: UTF-32 Little Endian (FF FE 00 00)")
+		return true
+	default:
+		log.Infoln("No BOM detected")
+		return false
 	}
+}
+
+func DropLastLine(b []byte) []byte {
 	for i := len(b) - 1; i > 0; i-- {
 		if b[i] == '\n' {
 			return b[:i]
@@ -27,7 +48,7 @@ var __obimimetype_registred__ = false
 func RegisterOBIMimeType() {
 	if !__obimimetype_registred__ {
 		csv := func(in []byte, limit uint32) bool {
-			in = DropLastLine(in, limit)
+			in = DropLastLine(in)
 
 			br := bytes.NewReader(in)
 			r := csv.NewReader(br)