| 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348 |
- // SPDX-FileCopyrightText: Copyright The Miniflux Authors. All rights reserved.
- // SPDX-License-Identifier: Apache-2.0
- package encoding // import "miniflux.app/v2/internal/reader/encoding"
- import (
- "bytes"
- "io"
- "os"
- "testing"
- "unicode/utf8"
- )
- func TestCharsetReaderWithUTF8(t *testing.T) {
- file := "testdata/utf8.xml"
- f, err := os.Open(file)
- if err != nil {
- t.Fatalf("Unable to open file: %v", err)
- }
- reader, err := CharsetReader("UTF-8", f)
- if err != nil {
- t.Fatalf("Unable to create reader: %v", err)
- }
- data, err := io.ReadAll(reader)
- if err != nil {
- t.Fatalf("Unable to read data: %v", err)
- }
- if !utf8.Valid(data) {
- t.Fatalf("Data is not valid UTF-8")
- }
- expectedUnicodeString := "Café"
- if !bytes.Contains(data, []byte(expectedUnicodeString)) {
- t.Fatalf("Data does not contain expected unicode string: %s", expectedUnicodeString)
- }
- }
- func TestCharsetReaderWithISO88591(t *testing.T) {
- file := "testdata/iso-8859-1.xml"
- f, err := os.Open(file)
- if err != nil {
- t.Fatalf("Unable to open file: %v", err)
- }
- reader, err := CharsetReader("ISO-8859-1", f)
- if err != nil {
- t.Fatalf("Unable to create reader: %v", err)
- }
- data, err := io.ReadAll(reader)
- if err != nil {
- t.Fatalf("Unable to read data: %v", err)
- }
- if !utf8.Valid(data) {
- t.Fatalf("Data is not valid UTF-8")
- }
- expectedUnicodeString := "Café"
- if !bytes.Contains(data, []byte(expectedUnicodeString)) {
- t.Fatalf("Data does not contain expected unicode string: %s", expectedUnicodeString)
- }
- }
- func TestCharsetReaderWithWindows1252(t *testing.T) {
- file := "testdata/windows-1252.xml"
- f, err := os.Open(file)
- if err != nil {
- t.Fatalf("Unable to open file: %v", err)
- }
- reader, err := CharsetReader("windows-1252", f)
- if err != nil {
- t.Fatalf("Unable to create reader: %v", err)
- }
- data, err := io.ReadAll(reader)
- if err != nil {
- t.Fatalf("Unable to read data: %v", err)
- }
- if !utf8.Valid(data) {
- t.Fatalf("Data is not valid UTF-8")
- }
- expectedUnicodeString := "Euro €"
- if !bytes.Contains(data, []byte(expectedUnicodeString)) {
- t.Fatalf("Data does not contain expected unicode string: %s", expectedUnicodeString)
- }
- }
- func TestCharsetReaderWithInvalidProlog(t *testing.T) {
- file := "testdata/invalid-prolog.xml"
- f, err := os.Open(file)
- if err != nil {
- t.Fatalf("Unable to open file: %v", err)
- }
- reader, err := CharsetReader("invalid", f)
- if err != nil {
- t.Fatalf("Unable to create reader: %v", err)
- }
- data, err := io.ReadAll(reader)
- if err != nil {
- t.Fatalf("Unable to read data: %v", err)
- }
- if !utf8.Valid(data) {
- t.Fatalf("Data is not valid UTF-8")
- }
- expectedUnicodeString := "Café"
- if !bytes.Contains(data, []byte(expectedUnicodeString)) {
- t.Fatalf("Data does not contain expected unicode string: %s", expectedUnicodeString)
- }
- }
- func TestCharsetReaderWithUTF8DocumentWithIncorrectProlog(t *testing.T) {
- file := "testdata/utf8-incorrect-prolog.xml"
- f, err := os.Open(file)
- if err != nil {
- t.Fatalf("Unable to open file: %v", err)
- }
- reader, err := CharsetReader("ISO-8859-1", f)
- if err != nil {
- t.Fatalf("Unable to create reader: %v", err)
- }
- data, err := io.ReadAll(reader)
- if err != nil {
- t.Fatalf("Unable to read data: %v", err)
- }
- if !utf8.Valid(data) {
- t.Fatalf("Data is not valid UTF-8")
- }
- expectedUnicodeString := "Café"
- if !bytes.Contains(data, []byte(expectedUnicodeString)) {
- t.Fatalf("Data does not contain expected unicode string: %s", expectedUnicodeString)
- }
- }
- func TestCharsetReaderWithWindows1252DocumentWithIncorrectProlog(t *testing.T) {
- file := "testdata/windows-1252-incorrect-prolog.xml"
- f, err := os.Open(file)
- if err != nil {
- t.Fatalf("Unable to open file: %v", err)
- }
- reader, err := CharsetReader("windows-1252", f)
- if err != nil {
- t.Fatalf("Unable to create reader: %v", err)
- }
- data, err := io.ReadAll(reader)
- if err != nil {
- t.Fatalf("Unable to read data: %v", err)
- }
- if !utf8.Valid(data) {
- t.Fatalf("Data is not valid UTF-8")
- }
- expectedUnicodeString := "Euro €"
- if !bytes.Contains(data, []byte(expectedUnicodeString)) {
- t.Fatalf("Data does not contain expected unicode string: %s", expectedUnicodeString)
- }
- }
- func TestNewReaderWithUTF8Document(t *testing.T) {
- file := "testdata/utf8.html"
- f, err := os.Open(file)
- if err != nil {
- t.Fatalf("Unable to open file: %v", err)
- }
- reader, err := NewCharsetReader(f, "text/html; charset=UTF-8")
- if err != nil {
- t.Fatalf("Unable to create reader: %v", err)
- }
- data, err := io.ReadAll(reader)
- if err != nil {
- t.Fatalf("Unable to read data: %v", err)
- }
- if !utf8.Valid(data) {
- t.Fatalf("Data is not valid UTF-8")
- }
- expectedUnicodeString := "Café"
- if !bytes.Contains(data, []byte(expectedUnicodeString)) {
- t.Fatalf("Data does not contain expected unicode string: %s", expectedUnicodeString)
- }
- }
- func TestNewReaderWithUTF8DocumentAndNoContentEncoding(t *testing.T) {
- file := "testdata/utf8.html"
- f, err := os.Open(file)
- if err != nil {
- t.Fatalf("Unable to open file: %v", err)
- }
- reader, err := NewCharsetReader(f, "text/html")
- if err != nil {
- t.Fatalf("Unable to create reader: %v", err)
- }
- data, err := io.ReadAll(reader)
- if err != nil {
- t.Fatalf("Unable to read data: %v", err)
- }
- if !utf8.Valid(data) {
- t.Fatalf("Data is not valid UTF-8")
- }
- expectedUnicodeString := "Café"
- if !bytes.Contains(data, []byte(expectedUnicodeString)) {
- t.Fatalf("Data does not contain expected unicode string: %s", expectedUnicodeString)
- }
- }
- func TestNewReaderWithISO88591Document(t *testing.T) {
- file := "testdata/iso-8859-1.xml"
- f, err := os.Open(file)
- if err != nil {
- t.Fatalf("Unable to open file: %v", err)
- }
- reader, err := NewCharsetReader(f, "text/html; charset=ISO-8859-1")
- if err != nil {
- t.Fatalf("Unable to create reader: %v", err)
- }
- data, err := io.ReadAll(reader)
- if err != nil {
- t.Fatalf("Unable to read data: %v", err)
- }
- if !utf8.Valid(data) {
- t.Fatalf("Data is not valid UTF-8")
- }
- expectedUnicodeString := "Café"
- if !bytes.Contains(data, []byte(expectedUnicodeString)) {
- t.Fatalf("Data does not contain expected unicode string: %s", expectedUnicodeString)
- }
- }
- func TestNewReaderWithISO88591DocumentAndNoContentType(t *testing.T) {
- file := "testdata/iso-8859-1.xml"
- f, err := os.Open(file)
- if err != nil {
- t.Fatalf("Unable to open file: %v", err)
- }
- reader, err := NewCharsetReader(f, "")
- if err != nil {
- t.Fatalf("Unable to create reader: %v", err)
- }
- data, err := io.ReadAll(reader)
- if err != nil {
- t.Fatalf("Unable to read data: %v", err)
- }
- if !utf8.Valid(data) {
- t.Fatalf("Data is not valid UTF-8")
- }
- expectedUnicodeString := "Café"
- if !bytes.Contains(data, []byte(expectedUnicodeString)) {
- t.Fatalf("Data does not contain expected unicode string: %s", expectedUnicodeString)
- }
- }
- func TestNewReaderWithISO88591DocumentWithMetaAfter1024Bytes(t *testing.T) {
- file := "testdata/iso-8859-1-meta-after-1024.html"
- f, err := os.Open(file)
- if err != nil {
- t.Fatalf("Unable to open file: %v", err)
- }
- reader, err := NewCharsetReader(f, "text/html")
- if err != nil {
- t.Fatalf("Unable to create reader: %v", err)
- }
- data, err := io.ReadAll(reader)
- if err != nil {
- t.Fatalf("Unable to read data: %v", err)
- }
- if !utf8.Valid(data) {
- t.Fatalf("Data is not valid UTF-8")
- }
- expectedUnicodeString := "Café"
- if !bytes.Contains(data, []byte(expectedUnicodeString)) {
- t.Fatalf("Data does not contain expected unicode string: %s", expectedUnicodeString)
- }
- }
- func TestNewReaderWithUTF8DocumentWithMetaAfter1024Bytes(t *testing.T) {
- file := "testdata/utf8-meta-after-1024.html"
- f, err := os.Open(file)
- if err != nil {
- t.Fatalf("Unable to open file: %v", err)
- }
- reader, err := NewCharsetReader(f, "text/html")
- if err != nil {
- t.Fatalf("Unable to create reader: %v", err)
- }
- data, err := io.ReadAll(reader)
- if err != nil {
- t.Fatalf("Unable to read data: %v", err)
- }
- if !utf8.Valid(data) {
- t.Fatalf("Data is not valid UTF-8")
- }
- expectedUnicodeString := "Café"
- if !bytes.Contains(data, []byte(expectedUnicodeString)) {
- t.Fatalf("Data does not contain expected unicode string: %s", expectedUnicodeString)
- }
- }
|