[JDEV] adding authentication associated with a tranport.
Oliver George
oliver at littledevil.com.au
Thu Oct 19 22:09:04 CDT 2000
to give you an idea, look at this diff...
oliver at devilmask:~/projects/jabber/jabber-1.1.3/pthsock$ diff client.c
flatclient.c
137a138
> io_write_null(cur);
173a175
> io_write_null(cur);
193c195,196
< io_write_str(c,xstream_header_char(h));
---
> io_write_str(c,xstream_flatheader_char(h));
> io_write_null(c);
240a244
> io_write_null(c);
280a285
> int index;
292c297,302
< ret=xstream_eat(c->xs,buffer,bufsz);
---
> index=0;
> while (index<bufsz) {
> if (buffer[index] != '\0')
> ret=xstream_eat(c->xs,&buffer[index],1);
> index++;
> }
and this is the addition to io_select.c
void io_write_null(sock c)
{
if(c->xbuffer!=NULL) _io_write_dump(c);
pth_write(c->fd,"\0",1);
}
these changes will be included with a namespace switch, initially i'm
just compiling as a separate .so and listening on a separate socket.
Hope this is useful. Oliver.
dlb wrote:
> congratulations on getting this to work !
> Have you implemented the revision to io_select's io_write_dump
> function which you had discussed - using a buffer length variable ?
> OR perhaps another means ??
>
> any advice you can provide will be greatly appreciated. I've heard
> that building and configuring jabber is no small feat, but I'll give
> it a shot if this work-around is effective.
>
> I'd been contemplating developing a socket interchange in TCL - to
> intermediate and filter the stream from F5- which would address this
> same problem, but your solution is preferable.
>
> nice work
>
>
> Oliver George wrote:
>
>> Hi JDEVers,
>>
>> (in case anyone is interested i have successfully interfaced flash 5
>> with
>> jabber and hopefully i will integrate these changes to the jabber source
>> - thanks to all who helped)
>>
>> <http://mailman.jabber.org/listinfo/jdev>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.jabber.org/jdev/attachments/20001020/343f3bd8/attachment-0002.htm>
More information about the JDev
mailing list