I receive from the server through curl the data in utf.
The server likes to shorten the text by simply cutting the string byte by bit, so the text is full at the ends of the line, after which there is a three-point
If I take this text through curl and give it to the regular, the regular will refuse to process it.
$ return = str_replace ("", "", $ return); did not remove problem areas
$ return = mb_convert_encoding ($ return, 'UTF-8', 'UTF-8'); partly helped, but now in the text began to slip conomic ئ nbsp; .., which also does not chew on regular basis