andyg;617467 Wrote: > On Mar 12, 2011, at 4:00 AM, pippin wrote: > > > Let's face it: the fundamental problem here (which you and I can't > > change right now, I'm aware of that) is that there is absolutely no > > documentation about how things _should_ work but we all - you, > > Logitech, me, other 3rd party developers - are fudging up a system > that > > somehow works by incidence and as soon as _anything_ changes it can > > break something. > > Well I would say the 7.5.4 situation is not typical, we had to add > long-polling comet support and it happened to break streaming comet for > a short time. And gzip was also added to improve efficiency (but this > shouldn't have broken anything). I'm actually surprised it causes > problems on iOS as I would hope the OS HTTP client would abstract that > all away for you. Actually that seems to work for me.
bluegaspode;617491 Wrote: > One cannot use the highlevel iOS networking classes as they sometimes > decide to cache the comet chunks and don't hand them over for > processing instantly. > So I'm using a low level HTTP-class and was not aware that it sends the > gzip headers but does not decompress automatically. Have a working > version now but need to wait 5-10 days until Apple reviews. Turning on > comet debug log helps because this avoid the compression code on the > sever side. Hm, I've got no problem with that, I use CFStream and it seems to handle the decompressionAre we talking about anything else but the initial headers anyway (I didn't check)? -- pippin --- see iPeng, the Squeezebox iPhone remote and *New: iPeng for iPad*, at penguinlovesmusic.com ------------------------------------------------------------------------ pippin's Profile: http://forums.slimdevices.com/member.php?userid=13777 View this thread: http://forums.slimdevices.com/showthread.php?t=86269 _______________________________________________ squeezenetwork mailing list [email protected] http://lists.slimdevices.com/mailman/listinfo/squeezenetwork
