[Libwebsockets] Segfault after doing lws_ring_insert

[prelag at yahoo.com]

You da man!






On Wednesday, August 5, 2020, 05:02:28 PM EDT, Paweł Konowalski <pawel.konowalski.delta at gmail.com> wrote: 





Paul,

Find the difference.
int len = strlen(response.c_str() + LWS_PRE);

vs

int len = strlen(response.c_str()) + LWS_PRE;
int len = response.length() + LWS_PRE;

BR
Paweł

W dniu 05.08.2020, śro o godzinie 19∶20 +0000, użytkownik
prelag at yahoo.com napisał:

> Thank you sir!
> 
> Am I supposed to set msg->first and msg->final?  If so, what are
> those set to?
> 
> Thanks
> 
> Paul R.
> 
> 
> 
> On Wednesday, August 5, 2020, 02:48:46 PM EDT, Andy Green <andy at warmc
> at.com> wrote: 
> 
> 
> 
> 
> 
> 
> 
> On 8/5/20 7:37 PM, prelag at yahoo.com wrote:
> > Hello,
> > 
> > Thanks in advance!  I know this is something simple I am
> > overlooking.
> > 
> > I have a test program based on the client echo example.  I am
> > attempting to create a message and insert it into the ring_buffer
> > to get picked up on the next LWS_CALLBACK_CLIENT_WRITEABLE.
> > 
> > Here is my simple function that is called when a message is ready
> > to be put in the buffer:
> > 
> > void MyClass::sendResponse(string &response) {
> >       syslog(LOG_DEBUG, "MyClass::sendResponse %s",
> > response.c_str());
> > 
> > 
> >       int len = strlen(response.c_str() + LWS_PRE);
> > 
> > 
> >       msg3.payload = malloc(len);
> 
> You're writing beyond what you allocated-->
> 
> 
> > 
> >       memcpy((char *) msg3.payload + LWS_PRE, response.c_str(),
> > len);
> 
> 
> None of that is any lws issue.
> 
> -Andy

> 
> _______________________________________________
> Libwebsockets mailing list
> Libwebsockets at ml.libwebsockets.org
> https://libwebsockets.org/mailman/listinfo/libwebsockets