Add php_stream_get_record_unicde() API call.

Update stream_get_line() userspace function to handle unicode streams.

3 files changed, 99 insertions, 9 deletions

diff --git a/ext/standard/streamsfuncs.c b/ext/standard/streamsfuncs.c
index 0cfd37e3de..ace8ce1e05 100644
--- a/ext/standard/streamsfuncs.c
+++ b/ext/standard/streamsfuncs.c
@@ -1235,18 +1235,15 @@ PHP_FUNCTION(stream_filter_remove)
 
 /* {{{ proto string stream_get_line(resource stream, int maxlen [, string ending])
    Read up to maxlen bytes from a stream or until the ending string is found */
-/* UTODO */
 PHP_FUNCTION(stream_get_line)
 {
-	char *str = NULL;
-	int str_len;
 	long max_length;
 	zval *zstream;
-	char *buf;
 	size_t buf_size;
 	php_stream *stream;
-	
-	if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "rl|s", &zstream, &max_length, &str, &str_len) == FAILURE) {
+	zval **delim = NULL;
+
+	if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "rl|Z", &zstream, &max_length, &delim) == FAILURE) {
 		RETURN_FALSE;
 	}
 
@@ -1260,10 +1257,34 @@ PHP_FUNCTION(stream_get_line)
 
 	php_stream_from_zval(stream, &zstream);
 
-	if ((buf = php_stream_get_record(stream, max_length, &buf_size, str, str_len TSRMLS_CC))) {
-		RETURN_STRINGL(buf, buf_size, 0);
+	if (php_stream_reads_unicode(stream)) {
+		UChar *buf;
+
+		if (Z_TYPE_PP(delim) != IS_UNICODE) {
+			convert_to_unicode_ex(delim);
+		}
+
+		/* maxchars == maxlength will prevent the otherwise generous maxlen == max_length * 2
+		   from allocating beyond what's requested */
+		buf = php_stream_get_record_unicode(stream, max_length * 2, max_length, &buf_size, Z_USTRVAL_PP(delim), Z_USTRLEN_PP(delim) TSRMLS_CC);
+		if (!buf) {
+			RETURN_FALSE;
+		}
+
+		RETURN_UNICODEL(buf, buf_size, 0);
 	} else {
-		RETURN_FALSE;
+		char *buf;
+
+		if (Z_TYPE_PP(delim) != IS_STRING) {
+			convert_to_string_ex(delim);
+		}
+
+		buf = php_stream_get_record(stream, max_length, &buf_size, Z_STRVAL_PP(delim), Z_STRLEN_PP(delim) TSRMLS_CC);
+		if (!buf) {
+			RETURN_FALSE;
+		}
+
+		RETURN_STRINGL(buf, buf_size, 0);
 	}
 }
 
diff --git a/main/php_streams.h b/main/php_streams.h
index 1bb4533443..d62a103937 100755
--- a/main/php_streams.h
+++ b/main/php_streams.h
@@ -329,6 +329,8 @@ PHPAPI void *_php_stream_get_line(php_stream *stream, int buf_type, zstr buf, si
 															_php_stream_get_line((stream), (buf_type), ZSTR(buf), (maxlen), (maxchars), NULL TSRMLS_CC)
 
 PHPAPI char *php_stream_get_record(php_stream *stream, size_t maxlen, size_t *returned_len, char *delim, size_t delim_len TSRMLS_DC);
+PHPAPI UChar *php_stream_get_record_unicode(php_stream *stream, size_t maxlen, size_t maxchars, size_t *returned_len, UChar *delim, size_t delim_len TSRMLS_DC);
+
 
 PHPAPI UChar *_php_stream_u_get_line(php_stream *stream, UChar *buf, int32_t *pmax_bytes, int32_t *pmax_chars, int *pis_unicode TSRMLS_DC);
 #define php_stream_u_get_line(stream, buf, maxlen_buf, maxlen_chars, buf_type)	_php_stream_u_get_line((stream), (buf), (maxlen_buf), (maxlen_chars), (buf_type) TSRMLS_CC)
diff --git a/main/streams/streams.c b/main/streams/streams.c
index 597ce80f38..be39d44270 100755
--- a/main/streams/streams.c
+++ b/main/streams/streams.c
@@ -1171,6 +1171,73 @@ PHPAPI char *php_stream_get_record(php_stream *stream, size_t maxlen, size_t *re
 	}
 }
 
+PHPAPI UChar *php_stream_get_record_unicode(php_stream *stream, size_t maxlen, size_t maxchars, size_t *returned_len, UChar *delim, size_t delim_len TSRMLS_DC)
+{
+	UChar *e, *buf;
+	size_t toread;
+	int skip = 0;
+
+	if (!php_stream_reads_unicode(stream)) {
+		return NULL;
+	}
+
+	php_stream_fill_read_buffer(stream, maxlen TSRMLS_CC);
+
+	if (delim_len == 0 || !delim) {
+		toread = maxlen;
+	} else {
+		if (delim_len == 1) {
+			e = u_memchr(stream->readbuf.u + stream->readpos, *delim, stream->writepos - stream->readpos);
+		} else {
+			e = u_strFindFirst(stream->readbuf.u + stream->readpos, stream->writepos - stream->readpos, delim, delim_len);
+		}
+
+		if (!e) {
+			toread = maxlen;
+		} else {
+			toread = e - (stream->readbuf.u + stream->readpos);
+			skip = 1;
+		}
+	}
+
+	if (toread > maxlen && maxlen > 0) {
+		toread = maxlen;
+	}
+
+	if (U16_IS_SURROGATE(stream->readbuf.u[stream->readpos + toread - 1]) &&
+		U16_IS_SURROGATE_LEAD(stream->readbuf.u[stream->readpos + toread - 1])) {
+		/* Don't orphan */
+		toread--;
+	}
+
+	if (maxchars > 0) {
+		size_t ulen = u_countChar32(stream->readbuf.u + stream->readpos, toread);
+
+		if (maxchars > ulen) {
+			int i = 0;
+			UChar *s = stream->readbuf.u + stream->readpos;
+
+			U16_FWD_N(s, i, toread, maxchars);
+			toread = i;
+		}
+	}
+
+	buf = eumalloc(toread + 1);
+	*returned_len = php_stream_read_unicode(stream, buf, toread);
+
+	if (*returned_len >= 0) {
+		if (skip) {
+			stream->readpos += delim_len;
+			stream->position += delim_len;
+		}
+		buf[*returned_len] = 0;
+		return buf;
+	} else {
+		efree(buf);
+		return NULL;
+	}
+}
+
 /* Writes a buffer directly to a stream, using multiple of the chunk size */
 static size_t _php_stream_write_buffer(php_stream *stream, int buf_type, zstr buf, int buflen TSRMLS_DC)
 {