Server can't receive non-ascii text correctly about emacs-websocket

[ahyatt]

I can reproduce this. I think I may have just fixed the client behavior, not the server behavior. This looks a bit tricky, but I'll take a look and see if I can fix it in the next week.

[alpha22jp]

Hi @ahyatt,

I found the cause. It's actually a problem of test code. The string server received is UTF-8 byte sequence, so it has to be converted to unibyte format before we pass it to websocket-send-text function.

After I modified test code as below, the test passed successfully.

--- old.el        2016-10-22 10:44:21.210792558 +0900
+++ new.el        2016-10-22 10:43:56.826793474 +0900
@@ -5,7 +5,7 @@
                     :on-message (lambda (ws frame)
                                   (message "Server received text!")
                                   (websocket-send-text ws
-                                                       (websocket-frame-payload frame)))
+                                                       (decode-coding-string (string-make-unibyte (websocket-frame-payload frame)) 'utf-8)))
                     :on-open (lambda (_websocket) "Client connection opened!")
                     :on-close (lambda (_websocket)
                                 (setq wstest-closed t)))))

[ahyatt]

That's interesting that you solved it this way.... I looked at this yesterday and traced this down to the websocket-mask function, which incorrectly makes a string of bytes in such a way that it is interpreted as a multibyte string of what are actually unibytes. So I think working around this problem in this way may work, but shouldn't be necessary. I'll take a look tomorrow about the best way to fix it - maybe just wrapping that function in string-make-unibyte would be best.

[ahyatt]

Should be fixed in f7d3fb5, if you use the handy new function websocket-frame-text, please re-open if not.

[ahyatt]

That's interesting that you solved it this way.... I looked at this
yesterday and traced this down to the websocket-mask function, which
incorrectly makes a string of bytes in such a way that it is interpreted as
a multibyte string of what are actually unibytes. So I think working
around this problem in this way may work, but shouldn't be necessary. I'll
take a look tomorrow about the best way to fix it.

On Fri, Oct 21, 2016 at 9:53 PM alpha22jp [email protected] wrote:

Hi @ahyatt https://github.com/ahyatt,

I found the cause. It's actually a problem of test code. The string server
received is UTF-8 byte sequence, so it has to be converted to unibyte
format before we pass it to websocket-send-text function.

After I modified test code as below, the test passed successfully.

--- old.el 2016-10-22 10:44:21.210792558 +0900+++ new.el 2016-10-22 10:43:56.826793474 +0900@@ -5,7 +5,7 @@

                 :on-message (lambda (ws frame)
                               (message "Server received text!")
                               (websocket-send-text ws

•                                                   (websocket-frame-payload frame)))+                                                       (decode-coding-string (string-make-unibyte (websocket-frame-payload frame)) 'utf-8)))
  
  
               :on-open (lambda (_websocket) "Client connection opened!")
               :on-close (lambda (_websocket)
                           (setq wstest-closed t)))))
  

—
You are receiving this because you were mentioned.

Reply to this email directly, view it on GitHub
#45 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AABP5LQpSYcnx1yjNLJw3lq5MWYZNRCsks5q2WyZgaJpZM4KaYFg
.