Merged revisions 76230 via svnmerge from

svn+ssh://pythondev@svn.python.org/python/trunk

........
  r76230 | benjamin.peterson | 2009-11-12 17:39:44 -0600 (Thu, 12 Nov 2009) | 2 lines

  fix several compile() issues by translating newlines in the tokenizer
........

3 files changed, 80 insertions, 22 deletions

diff --git a/Parser/parsetok.c b/Parser/parsetok.c
index 90a90a7..ff4ca70 100644
--- a/Parser/parsetok.c
+++ b/Parser/parsetok.c
@@ -46,13 +46,14 @@ PyParser_ParseStringFlagsFilenameEx(const char *s, const char *filename,
 		          perrdetail *err_ret, int *flags)
 {
 	struct tok_state *tok;
+	int exec_input = start == file_input;
 
 	initerr(err_ret, filename);
 
 	if (*flags & PyPARSE_IGNORE_COOKIE)
-		tok = PyTokenizer_FromUTF8(s);
+		tok = PyTokenizer_FromUTF8(s, exec_input);
 	else
-		tok = PyTokenizer_FromString(s);
+		tok = PyTokenizer_FromString(s, exec_input);
 	if (tok == NULL) {
 		err_ret->error = PyErr_Occurred() ? E_DECODE : E_NOMEM;
 		return NULL;
diff --git a/Parser/tokenizer.c b/Parser/tokenizer.c
index daf18dc..e637cb3 100644
--- a/Parser/tokenizer.c
+++ b/Parser/tokenizer.c
@@ -119,6 +119,7 @@ tok_new(void)
 	tok->buf = tok->cur = tok->end = tok->inp = tok->start = NULL;
 	tok->done = E_OK;
 	tok->fp = NULL;
+	tok->input = NULL;
 	tok->tabsize = TABSIZE;
 	tok->indent = 0;
 	tok->indstack[0] = 0;
@@ -145,6 +146,17 @@ tok_new(void)
 	return tok;
 }
 
+static char *
+new_string(const char *s, Py_ssize_t len)
+{
+	char* result = (char *)PyMem_MALLOC(len + 1);
+	if (result != NULL) {
+		memcpy(result, s, len);
+		result[len] = '\0';
+	}
+	return result;
+}
+
 #ifdef PGEN
 
 static char *
@@ -159,10 +171,10 @@ decoding_feof(struct tok_state *tok)
 	return feof(tok->fp);
 }
 
-static const char *
-decode_str(const char *str, struct tok_state *tok)
+static char *
+decode_str(const char *str, int exec_input, struct tok_state *tok)
 {
-	return str;
+	return new_string(str, strlen(str));
 }
 
 #else /* PGEN */
@@ -177,16 +189,6 @@ error_ret(struct tok_state *tok) /* XXX */
 	return NULL;		/* as if it were EOF */
 }
 
-static char *
-new_string(const char *s, Py_ssize_t len)
-{
-	char* result = (char *)PyMem_MALLOC(len + 1);
-	if (result != NULL) {
-		memcpy(result, s, len);
-		result[len] = '\0';
-	}
-	return result;
-}
 
 static char *
 get_normal_name(char *s)	/* for utf-8 and latin-1 */
@@ -635,17 +637,63 @@ translate_into_utf8(const char* str, const char* enc) {
 	return utf8;
 }
 
+
+static char *
+translate_newlines(const char *s, int exec_input, struct tok_state *tok) {
+	int skip_next_lf = 0, length = strlen(s), final_length;
+	char *buf, *current;
+	char c;
+	buf = PyMem_MALLOC(length + 2);
+	if (buf == NULL) {
+		tok->done = E_NOMEM;
+		return NULL;
+	}
+	for (current = buf; (c = *s++);) {
+		if (skip_next_lf) {
+			skip_next_lf = 0;
+			if (c == '\n') {
+				c = *s;
+				s++;
+				if (!c)
+					break;
+			}
+		}
+		if (c == '\r') {
+			skip_next_lf = 1;
+			c = '\n';
+		}
+		*current = c;
+		current++;
+	}
+	/* If this is exec input, add a newline to the end of the file if
+	   there isn't one already. */
+	if (exec_input && *current != '\n') {
+		*current = '\n';
+		current++;
+	}
+	*current = '\0';
+	final_length = current - buf;
+	if (final_length < length && final_length)
+		/* should never fail */
+		buf = PyMem_REALLOC(buf, final_length + 1);
+	return buf;
+}
+
 /* Decode a byte string STR for use as the buffer of TOK.
    Look for encoding declarations inside STR, and record them
    inside TOK.  */
 
 static const char *
-decode_str(const char *str, struct tok_state *tok)
+decode_str(const char *input, int single, struct tok_state *tok)
 {
 	PyObject* utf8 = NULL;
+	const char *str;
 	const char *s;
 	const char *newl[2] = {NULL, NULL};
 	int lineno = 0;
+	tok->input = str = translate_newlines(input, single, tok);
+	if (str == NULL)
+		return NULL;
 	tok->enc = NULL;
 	tok->str = str;
 	if (!check_bom(buf_getc, buf_ungetc, buf_setreadl, tok))
@@ -696,12 +744,12 @@ decode_str(const char *str, struct tok_state *tok)
 /* Set up tokenizer for string */
 
 struct tok_state *
-PyTokenizer_FromString(const char *str)
+PyTokenizer_FromString(const char *str, int exec_input)
 {
 	struct tok_state *tok = tok_new();
 	if (tok == NULL)
 		return NULL;
-	str = (char *)decode_str(str, tok);
+	str = (char *)decode_str(str, exec_input, tok);
 	if (str == NULL) {
 		PyTokenizer_Free(tok);
 		return NULL;
@@ -713,11 +761,18 @@ PyTokenizer_FromString(const char *str)
 }
 
 struct tok_state *
-PyTokenizer_FromUTF8(const char *str)
+PyTokenizer_FromUTF8(const char *str, int exec_input)
 {
 	struct tok_state *tok = tok_new();
 	if (tok == NULL)
 		return NULL;
+#ifndef PGEN
+	tok->input = str = translate_newlines(str, exec_input, tok);
+#endif
+	if (str == NULL) {
+		PyTokenizer_Free(tok);
+		return NULL;
+	}
 	tok->decoding_state = STATE_RAW;
 	tok->read_coding_spec = 1;
 	tok->enc = NULL;
@@ -734,7 +789,6 @@ PyTokenizer_FromUTF8(const char *str)
 	return tok;
 }
 
-
 /* Set up tokenizer for file */
 
 struct tok_state *
@@ -780,6 +834,8 @@ PyTokenizer_Free(struct tok_state *tok)
 #endif
 	if (tok->fp != NULL && tok->buf != NULL)
 		PyMem_FREE(tok->buf);
+	if (tok->input)
+		PyMem_FREE((char *)tok->input);
 	PyMem_FREE(tok);
 }
 
diff --git a/Parser/tokenizer.h b/Parser/tokenizer.h
index e3328f1..1a81e33 100644
--- a/Parser/tokenizer.h
+++ b/Parser/tokenizer.h
@@ -58,10 +58,11 @@ struct tok_state {
 #endif
 	const char* enc;        /* Encoding for the current str. */
 	const char* str;
+	const char* input; /* Tokenizer's newline translated copy of the string. */
 };
 
-extern struct tok_state *PyTokenizer_FromString(const char *);
-extern struct tok_state *PyTokenizer_FromUTF8(const char *);
+extern struct tok_state *PyTokenizer_FromString(const char *, int);
+extern struct tok_state *PyTokenizer_FromUTF8(const char *, int);
 extern struct tok_state *PyTokenizer_FromFile(FILE *, char*,
 					      char *, char *);
 extern void PyTokenizer_Free(struct tok_state *);