Latin-1 source code was not being properly decoded when passed through

compile(). This was due to left-over special-casing before UTF-8 became the
default source encoding.

Closes issue #3574. Thanks to Victor Stinner for help with the patch.

2 files changed, 4 insertions, 4 deletions

diff --git a/Parser/tokenizer.c b/Parser/tokenizer.c
index 4edf6d0..ce8129d 100644
--- a/Parser/tokenizer.c
+++ b/Parser/tokenizer.c
@@ -135,6 +135,7 @@ tok_new(void)
 	tok->decoding_state = STATE_INIT;
 	tok->decoding_erred = 0;
 	tok->read_coding_spec = 0;
+	tok->enc = NULL;
 	tok->encoding = NULL;
         tok->cont_line = 0;
 #ifndef PGEN
@@ -274,8 +275,7 @@ check_coding_spec(const char* line, Py_ssize_t size, struct tok_state *tok,
 		tok->read_coding_spec = 1;
 		if (tok->encoding == NULL) {
 			assert(tok->decoding_state == STATE_RAW);
-			if (strcmp(cs, "utf-8") == 0 ||
-			    strcmp(cs, "iso-8859-1") == 0) {
+			if (strcmp(cs, "utf-8") == 0) {
 				tok->encoding = cs;
 			} else {
 				r = set_readline(tok, cs);
diff --git a/Parser/tokenizer.h b/Parser/tokenizer.h
index c45dea1..df9cbc7 100644
--- a/Parser/tokenizer.h
+++ b/Parser/tokenizer.h
@@ -49,14 +49,14 @@ struct tok_state {
 	enum decoding_state decoding_state;
 	int decoding_erred;	/* whether erred in decoding  */
 	int read_coding_spec;	/* whether 'coding:...' has been read  */
-	char *encoding;
+	char *encoding;         /* Source encoding. */
 	int cont_line;          /* whether we are in a continuation line. */
 	const char* line_start;	/* pointer to start of current line */
 #ifndef PGEN
 	PyObject *decoding_readline; /* codecs.open(...).readline */
 	PyObject *decoding_buffer;
 #endif
-	const char* enc;
+	const char* enc;        /* Encoding for the current str. */
 	const char* str;
 };