Merge pull request #63733 from akien-mga/file-get_as_text-skip-CR

File: Re-add support to skip CR (`\r`) in `File::get_as_text`

5 files changed, 46 insertions, 0 deletions

diff --git a/tests/core/io/test_file_access.h b/tests/core/io/test_file_access.h
index f0e1cceacf..aab62955cb 100644
--- a/tests/core/io/test_file_access.h
+++ b/tests/core/io/test_file_access.h
@@ -78,6 +78,29 @@ TEST_CASE("[FileAccess] CSV read") {
 	CHECK(row5[1] == "tab separated");
 	CHECK(row5[2] == "lines, good?");
 }
+
+TEST_CASE("[FileAccess] Get as UTF-8 String") {
+	Ref<FileAccess> f_lf = FileAccess::open(TestUtils::get_data_path("line_endings_lf.test.txt"), FileAccess::READ);
+	String s_lf = f_lf->get_as_utf8_string();
+	f_lf->seek(0);
+	String s_lf_nocr = f_lf->get_as_utf8_string(true);
+	CHECK(s_lf == "Hello darkness\nMy old friend\nI've come to talk\nWith you again\n");
+	CHECK(s_lf_nocr == "Hello darkness\nMy old friend\nI've come to talk\nWith you again\n");
+
+	Ref<FileAccess> f_crlf = FileAccess::open(TestUtils::get_data_path("line_endings_crlf.test.txt"), FileAccess::READ);
+	String s_crlf = f_crlf->get_as_utf8_string();
+	f_crlf->seek(0);
+	String s_crlf_nocr = f_crlf->get_as_utf8_string(true);
+	CHECK(s_crlf == "Hello darkness\r\nMy old friend\r\nI've come to talk\r\nWith you again\r\n");
+	CHECK(s_crlf_nocr == "Hello darkness\nMy old friend\nI've come to talk\nWith you again\n");
+
+	Ref<FileAccess> f_cr = FileAccess::open(TestUtils::get_data_path("line_endings_cr.test.txt"), FileAccess::READ);
+	String s_cr = f_cr->get_as_utf8_string();
+	f_cr->seek(0);
+	String s_cr_nocr = f_cr->get_as_utf8_string(true);
+	CHECK(s_cr == "Hello darkness\rMy old friend\rI've come to talk\rWith you again\r");
+	CHECK(s_cr_nocr == "Hello darknessMy old friendI've come to talkWith you again");
+}
 } // namespace TestFileAccess
 
 #endif // TEST_FILE_ACCESS_H
diff --git a/tests/core/string/test_string.h b/tests/core/string/test_string.h
index 0c5704d6c9..b8b766023a 100644
--- a/tests/core/string/test_string.h
+++ b/tests/core/string/test_string.h
@@ -152,6 +152,20 @@ TEST_CASE("[String] UTF16 with BOM") {
 	CHECK(String::utf16(cs) == s);
 }
 
+TEST_CASE("[String] UTF8 with CR") {
+	const String base = U"Hello darkness\r\nMy old friend\nI've come to talk\rWith you again";
+
+	String keep_cr;
+	Error err = keep_cr.parse_utf8(base.utf8().get_data());
+	CHECK(err == OK);
+	CHECK(keep_cr == base);
+
+	String no_cr;
+	err = no_cr.parse_utf8(base.utf8().get_data(), -1, true); // Skip CR.
+	CHECK(err == OK);
+	CHECK(no_cr == base.replace("\r", ""));
+}
+
 TEST_CASE("[String] Invalid UTF8 (non-standard)") {
 	ERR_PRINT_OFF
 	static const uint8_t u8str[] = { 0x45, 0xE3, 0x81, 0x8A, 0xE3, 0x82, 0x88, 0xE3, 0x81, 0x86, 0xF0, 0x9F, 0x8E, 0xA4, 0xF0, 0x82, 0x82, 0xAC, 0xED, 0xA0, 0x81, 0 };
diff --git a/tests/data/line_endings_cr.test.txt b/tests/data/line_endings_cr.test.txt
new file mode 100644
index 0000000000..556154bb25
--- /dev/null
+++ b/tests/data/line_endings_cr.test.txt
@@ -0,0 +1 @@
+Hello darkness
My old friend
I've come to talk
With you again
\ No newline at end of file
diff --git a/tests/data/line_endings_crlf.test.txt b/tests/data/line_endings_crlf.test.txt
new file mode 100644
index 0000000000..a3cbe55b7f
--- /dev/null
+++ b/tests/data/line_endings_crlf.test.txt
@@ -0,0 +1,4 @@
+Hello darkness

+My old friend

+I've come to talk

+With you again

diff --git a/tests/data/line_endings_lf.test.txt b/tests/data/line_endings_lf.test.txt
new file mode 100644
index 0000000000..0aabcd911e
--- /dev/null
+++ b/tests/data/line_endings_lf.test.txt
@@ -0,0 +1,4 @@
+Hello darkness
+My old friend
+I've come to talk
+With you again