如何在C中使用zlib库对websocket permessage-deflate进行编码和解码？

Question

All I am trying to encode WebSocket payload using per-message-deflate. I am trying to use some test code but my encoding is not proper can you please help me.

Here is my code:

char a[180] = "{\"type\":\"message\",\"channel\":\"C04U0F9CW\",\"text\":\"TestingMessage\",\"id\":757}";
char b[180];
char c[180];
int i;
// deflate
// zlib struct
z_stream defstream;
defstream.zalloc = Z_NULL;
defstream.zfree = Z_NULL;
defstream.opaque = Z_NULL;
defstream.avail_in = (uInt)strlen(a)+1; // size of input, string + terminator
defstream.next_in = (Bytef *)a; // input char array
defstream.avail_out = (uInt)sizeof(b); // size of output
defstream.next_out = (Bytef *)b; // output char array

//deflateInit(&defstream, Z_SYNC_FLUSH);
deflateInit2(&defstream, Z_DEFAULT_COMPRESSION, Z_DEFLATED, 15, 8, Z_DEFAULT_STRATEGY);
deflate(&defstream, Z_SYNC_FLUSH);
deflateEnd(&defstream);
for(i=0;i<=(char*)defstream.next_out - b;i++)
{
        printf("%X",(Bytef)b[i]);
}
printf("\n%s\n",(Bytef *)b);
// This is one way of getting the size of the output
printf("Deflated size is: %lu\n", (char*)defstream.next_out - b);

my output is

kml@kml-VirtualBox:~/zlib-1.2.11/test$ gcc test.c -lz && ./a.out 
789CAA562AA92C4855B252CA4D2D2E4E4C4F55D2514ACE48CCCB4BCD18A391B98841AB8593A873454B522B4A804221A9C5259979E9BE70D599294A56E6A6E6B5C0000FFFF2E
x��V*�,HU�R�M-.NLOU�QJ�H��K��9���Y:�EKR+J�B!��%�y��pՙ)JV����

Deflated size is: 72
Inflate:
73
{"type":"message","channel":"C04U0F9CW","text":"TestingMessage","id":757}

So I am expecting payload will me like this

AA562AA92C4855B252CA4D2D2E4E4C4F55D2514ACE48CCCB4BCD018A391B98841AB8593A8703454B522B4A804221A9C5259979E9BE70D599294A56E6A6E6B50000

This is the fiddler WebSocket traffic capture for that particular message

Please help me out to get proper per-message-deflate encoding.

Here is some more upgrade information
GET https://127.0.0.1/websoc.html HTTP/1.1
Host: mpmulti-yklu.slack-msgs.com
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:53.0) Gecko/20100101 Firefox/53.0
Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8
Accept-Language: en-US,en;q=0.5
Accept-Encoding: gzip, deflate, br
Sec-WebSocket-Version: 13
Sec-WebSocket-Extensions: permessage-deflate
Sec-WebSocket-Key: FriWg0AWhiLhLXXnfJK8kA==
Connection: keep-alive, Upgrade
Pragma: no-cache
Cache-Control: no-cache
Upgrade: websocket

Response

HTTP/1.1 101 Switching Protocols
Upgrade: websocket
Connection: Upgrade
Sec-WebSocket-Accept: S2Q+O/tRM0/mIG//Er2hDdqD0eY=
Sec-Websocket-Extensions: permessage-deflate; server_no_context_takeover; client_no_context_takeover
X-Via: haproxy-ext90

Answer 1

The code is correct but defstream.avail_in expect exact length, not for the extra null character. If you will provide exact input length it will work properly. Try like this

int l = (uInt)strlen(a);
defstream.avail_in = l; // size of input, string + terminator

Please try to understand how z_stream is assigning value check the code then you will better understand about this.

I think many people are trying the decode and they are not getting proper value for per-message-deflate.

Check your output

" 789CAA562AA92C4855B252CA4D2D2E4E4C4F55D2514ACE48CCCB4BCD18A391B98841AB8593A873454B522B4A804221A9C5259979E9BE70D599294A56E6A6E6B5C0000FFFF2E "

Here 789C is header and 0000FFFF or according to your code (00FFFF) is the trailer of deflate message. here 0 is for 00 because your output is print one '0' instead of '00'. you are getting 5C extra because of extra one size. Thye '2E' will also will be removed when you will try below code

int d = memcmp(msgend, &b[defstream.total_out - 4], 4) ? 0 : 4;
for(i=0;i<defstream.total_out - d;i++)
{
        printf("%2X",(Bytef)b[i]);
}

So your expected answer is

" AA562AA92C4855B252CA4D2D2E4E4C4F55D2514ACE48CCCB4BCD018A391B98841AB8593A8703454B522B4A804221A9C5259979E9BE70D599294A56E6A6E6B50000 "

As per your output, 2 bytes are 0 so as per your code it will be '00' and your output follow printing each nibble (half byte).

Please try this is worked for me and I am now able to decode properly.