support utf-32 file encoding

pkg/converters/custom/custom_transaction_data_dsv_file_importer.go

@@ -14,6 +14,7 @@ import (
 	"golang.org/x/text/encoding/simplifiedchinese"
 	"golang.org/x/text/encoding/traditionalchinese"
 	"golang.org/x/text/encoding/unicode"
+	"golang.org/x/text/encoding/unicode/utf32"
 	"golang.org/x/text/transform"
 
 	"github.com/mayswind/ezbookkeeping/pkg/converters/converter"
@@ -36,6 +37,8 @@ var supportedFileEncodings = map[string]encoding.Encoding{
 	"utf-8":        unicode.UTF8BOM,                                     // UTF-8
 	"utf-16le":     unicode.UTF16(unicode.LittleEndian, unicode.UseBOM), // UTF-16 Little Endian
 	"utf-16be":     unicode.UTF16(unicode.BigEndian, unicode.UseBOM),    // UTF-16 Big Endian
+	"utf-32le":     utf32.UTF32(utf32.LittleEndian, utf32.UseBOM),       // UTF-32 Little Endian
+	"utf-32be":     utf32.UTF32(utf32.BigEndian, utf32.UseBOM),          // UTF-32 Big Endian
 	"cp437":        charmap.CodePage437,                                 // OEM United States (CP-437)
 	"cp863":        charmap.CodePage863,                                 // OEM Canadian French (CP-863)
 	"cp037":        charmap.CodePage037,                                 // IBM EBCDIC US/Canada (CP-037)

src/consts/file.ts

@@ -15,6 +15,8 @@ export const SUPPORTED_FILE_ENCODINGS: string[] = [
     UTF_8, // UTF-8
     'utf-16le', // UTF-16 Little Endian
     'utf-16be', // UTF-16 Big Endian
+    'utf-32le', // UTF-32 Little Endian
+    'utf-32be', // UTF-32 Big Endian
     'cp437', // OEM United States (CP-437)
     'cp863', // OEM Canadian French (CP-863)
     'cp037', // IBM EBCDIC US/Canada (CP-037)
@@ -67,8 +69,8 @@ export const CHARDET_ENCODING_NAME_MAPPING: Record<string, string> = {
     'UTF-8': UTF_8,
     'UTF-16LE': 'utf-16le',
     'UTF-16BE': 'utf-16be',
-    // 'UTF-32 LE': '', // not supported
-    // 'UTF-32 BE': '', // not supported
+    'UTF-32LE': 'utf-32le',
+    'UTF-32BE': 'utf-32be',
     'ISO-2022-JP': 'iso-2022-jp',
     // 'ISO-2022-KR': '', // not supported
     // 'ISO-2022-CN': '', // not supported

src/locales/de.json

@@ -1355,6 +1355,8 @@
         "utf-8": "UTF-8",
         "utf-16le": "UTF-16 Little Endian",
         "utf-16be": "UTF-16 Big Endian",
+        "utf-32le": "UTF-32 Little Endian",
+        "utf-32be": "UTF-32 Big Endian",
         "cp437": "OEM United States (CP-437)",
         "cp863": "OEM Canadian French (CP-863)",
         "cp037": "IBM EBCDIC US/Canada (CP-037)",

src/locales/en.json

@@ -1355,6 +1355,8 @@
         "utf-8": "UTF-8",
         "utf-16le": "UTF-16 Little Endian",
         "utf-16be": "UTF-16 Big Endian",
+        "utf-32le": "UTF-32 Little Endian",
+        "utf-32be": "UTF-32 Big Endian",
         "cp437": "OEM United States (CP-437)",
         "cp863": "OEM Canadian French (CP-863)",
         "cp037": "IBM EBCDIC US/Canada (CP-037)",

src/locales/es.json

@@ -1355,6 +1355,8 @@
         "utf-8": "UTF-8",
         "utf-16le": "UTF-16 Little Endian",
         "utf-16be": "UTF-16 Big Endian",
+        "utf-32le": "UTF-32 Little Endian",
+        "utf-32be": "UTF-32 Big Endian",
         "cp437": "OEM United States (CP-437)",
         "cp863": "OEM Canadian French (CP-863)",
         "cp037": "IBM EBCDIC US/Canada (CP-037)",

src/locales/fr.json

@@ -1355,6 +1355,8 @@
         "utf-8": "UTF-8",
         "utf-16le": "UTF-16 Little Endian",
         "utf-16be": "UTF-16 Big Endian",
+        "utf-32le": "UTF-32 Little Endian",
+        "utf-32be": "UTF-32 Big Endian",
         "cp437": "OEM États-Unis (CP-437)",
         "cp863": "OEM Canadien Français (CP-863)",
         "cp037": "IBM EBCDIC États-Unis/Canada (CP-037)",

src/locales/it.json

@@ -1355,6 +1355,8 @@
         "utf-8": "UTF-8",
         "utf-16le": "UTF-16 Little Endian",
         "utf-16be": "UTF-16 Big Endian",
+        "utf-32le": "UTF-32 Little Endian",
+        "utf-32be": "UTF-32 Big Endian",
         "cp437": "OEM United States (CP-437)",
         "cp863": "OEM Canadian French (CP-863)",
         "cp037": "IBM EBCDIC US/Canada (CP-037)",

src/locales/ja.json

@@ -1355,6 +1355,8 @@
         "utf-8": "UTF-8",
         "utf-16le": "UTF-16 Little Endian",
         "utf-16be": "UTF-16 Big Endian",
+        "utf-32le": "UTF-32 Little Endian",
+        "utf-32be": "UTF-32 Big Endian",
         "cp437": "OEM 米国 (CP-437)",
         "cp863": "OEM カナダ系フランス語 (CP-863)",
         "cp037": "IBM EBCDIC 米国/カナダ (CP-037)",

src/locales/kn.json

@@ -1355,6 +1355,8 @@
         "utf-8": "UTF-8",
         "utf-16le": "UTF-16 ಲಿಟಲ್ ಎಂಡಿಯನ್",
         "utf-16be": "UTF-16 ಬಿಗ್ ಎಂಡಿಯನ್",
+        "utf-32le": "UTF-32 ಲಿಟಲ್ ಎಂಡಿಯನ್",
+        "utf-32be": "UTF-32 ಬಿಗ್ ಎಂಡಿಯನ್",
         "cp437": "OEM ಯುನೈಟೆಡ್ ಸ್ಟೇಟ್ಸ್ (CP-437)",
         "cp863": "OEM ಕ್ಯಾನಡಿಯನ್ ಫ್ರೆಂಚ್ (CP-863)",
         "cp037": "IBM EBCDIC ಯುಎಸ್/ಕ್ಯಾನಡಾ (CP-037)",

src/locales/ko.json

@@ -1355,6 +1355,8 @@
         "utf-8": "UTF-8",
         "utf-16le": "UTF-16 리틀 엔디안",
         "utf-16be": "UTF-16 빅 엔디안",
+        "utf-32le": "UTF-32 리틀 엔디안",
+        "utf-32be": "UTF-32 빅 엔디안",
         "cp437": "OEM 미국 (CP-437)",
         "cp863": "OEM 캐나다 프랑스어 (CP-863)",
         "cp037": "IBM EBCDIC 미국/캐나다 (CP-037)",

src/locales/nl.json

@@ -1355,6 +1355,8 @@
         "utf-8": "UTF-8",
         "utf-16le": "UTF-16 Little Endian",
         "utf-16be": "UTF-16 Big Endian",
+        "utf-32le": "UTF-32 Little Endian",
+        "utf-32be": "UTF-32 Big Endian",
         "cp437": "OEM United States (CP-437)",
         "cp863": "OEM Canadian French (CP-863)",
         "cp037": "IBM EBCDIC US/Canada (CP-037)",

src/locales/pt_BR.json

@@ -1355,6 +1355,8 @@
         "utf-8": "UTF-8",
         "utf-16le": "UTF-16 Little Endian",
         "utf-16be": "UTF-16 Big Endian",
+        "utf-32le": "UTF-32 Little Endian",
+        "utf-32be": "UTF-32 Big Endian",
         "cp437": "OEM Estados Unidos (CP-437)",
         "cp863": "OEM Francês Canadense (CP-863)",
         "cp037": "IBM EBCDIC EUA/Canadá (CP-037)",

src/locales/ru.json

@@ -1355,6 +1355,8 @@
         "utf-8": "UTF-8",
         "utf-16le": "UTF-16 Little Endian",
         "utf-16be": "UTF-16 Big Endian",
+        "utf-32le": "UTF-32 Little Endian",
+        "utf-32be": "UTF-32 Big Endian",
         "cp437": "OEM United States (CP-437)",
         "cp863": "OEM Canadian French (CP-863)",
         "cp037": "IBM EBCDIC US/Canada (CP-037)",

src/locales/sl.json

@@ -1355,6 +1355,8 @@
         "utf-8": "UTF-8",
         "utf-16le": "UTF-16 Little Endian",
         "utf-16be": "UTF-16 Big Endian",
+        "utf-32le": "UTF-32 Little Endian",
+        "utf-32be": "UTF-32 Big Endian",
         "cp437": "OEM Združene države (CP-437)",
         "cp863": "OEM kanadska francoščina (CP-863)",
         "cp037": "IBM EBCDIC ZDA/Kanada (CP-037)",

src/locales/ta.json

@@ -1355,6 +1355,8 @@
         "utf-8": "UTF-8",
         "utf-16le": "UTF-16 லிட்டில் எண்டியன்",
         "utf-16be": "UTF-16 பிக் எண்டியன்",
+        "utf-32le": "UTF-32 லிட்டில் எண்டியன்",
+        "utf-32be": "UTF-32 பிக் எண்டியன்",
         "cp437": "OEM அமெரிக்கா (CP-437)",
         "cp863": "OEM கனடா பிரஞ்சு (CP-863)",
         "cp037": "IBM EBCDIC அமெரிக்கா/கனடா (CP-037)",

src/locales/th.json

@@ -1355,6 +1355,8 @@
         "utf-8": "UTF-8",
         "utf-16le": "UTF-16 Endian เล็ก",
         "utf-16be": "UTF-16 Endian ใหญ่",
+        "utf-32le": "UTF-32 Endian เล็ก",
+        "utf-32be": "UTF-32 Endian ใหญ่",
         "cp437": "OEM สหรัฐอเมริกา (CP-437)",
         "cp863": "OEM ฝรั่งเศสแคนาดา (CP-863)",
         "cp037": "IBM EBCDIC สหรัฐอเมริกา/แคนาดา (CP-037)",

src/locales/tr.json

@@ -1355,6 +1355,8 @@
         "utf-8": "UTF-8",
         "utf-16le": "UTF-16 Little Endian",
         "utf-16be": "UTF-16 Big Endian",
+        "utf-32le": "UTF-32 Little Endian",
+        "utf-32be": "UTF-32 Big Endian",
         "cp437": "OEM Birleşik Devletler (CP-437)",
         "cp863": "OEM Kanada Fransızcası (CP-863)",
         "cp037": "IBM EBCDIC ABD/Kanada (CP-037)",

src/locales/uk.json

@@ -1355,6 +1355,8 @@
         "utf-8": "UTF-8",
         "utf-16le": "UTF-16 Little Endian",
         "utf-16be": "UTF-16 Big Endian",
+        "utf-32le": "UTF-32 Little Endian",
+        "utf-32be": "UTF-32 Big Endian",
         "cp437": "OEM United States (CP-437)",
         "cp863": "OEM Canadian French (CP-863)",
         "cp037": "IBM EBCDIC US/Canada (CP-037)",

src/locales/vi.json

@@ -1355,6 +1355,8 @@
         "utf-8": "UTF-8",
         "utf-16le": "UTF-16 Little Endian",
         "utf-16be": "UTF-16 Big Endian",
+        "utf-32le": "UTF-32 Little Endian",
+        "utf-32be": "UTF-32 Big Endian",
         "cp437": "OEM United States (CP-437)",
         "cp863": "OEM Canadian French (CP-863)",
         "cp037": "IBM EBCDIC US/Canada (CP-037)",

src/locales/zh_Hans.json

@@ -1355,6 +1355,8 @@
         "utf-8": "UTF-8",
         "utf-16le": "UTF-16 Little Endian",
         "utf-16be": "UTF-16 Big Endian",
+        "utf-32le": "UTF-32 Little Endian",
+        "utf-32be": "UTF-32 Big Endian",
         "cp437": "OEM 美国 (CP-437)",
         "cp863": "OEM 加拿大法语 (CP-863)",
         "cp037": "IBM EBCDIC 美国/加拿大 (CP-037)",

src/locales/zh_Hant.json

@@ -1355,6 +1355,8 @@
         "utf-8": "UTF-8",
         "utf-16le": "UTF-16 Little Endian",
         "utf-16be": "UTF-16 Big Endian",
+        "utf-32le": "UTF-32 Little Endian",
+        "utf-32be": "UTF-32 Big Endian",
         "cp437": "OEM 美國 (CP-437)",
         "cp863": "OEM 加拿大法語 (CP-863)",
         "cp037": "IBM EBCDIC 美國/加拿大 (CP-037)",