Thanks, Andrew, I'd hope by now you could give me a bit more credit than that. 
:)

The logs from CBL are quite clear. The reachability check is failing, but the 
server is definitely reachable using Safari - I can hit URLs on it and see the 
expected response. CBL is very clearly not ever attempting a connection due to 
the reachability check returning a failure. So, I am quite positive that in 
this situation, a negative response from SCNetworkReachability is NOT a clear 
indication that a successful request cannot be made, because the proxy server 
will allow it to succeed.

I do not particularly care WHY it reports a failure, nor was I attempting to 
posit a particular reason. In this environment, the server actually cannot be 
resolved by DNS, but again, that is irrelevant because it can connect to the 
proxy server, and only the proxy server needs to resolve the server.

So, I'm quite positive that when a proxy server is involved, 
SCNetworkReachability does not accurately predict whether or not a successful 
request to the server can be made. So, lets assume for the moment that I'm not 
an idiot and that I've diagnosed the issue correctly. In that case, do we think 
there are other likely problems with this? Seems like an actual check is the 
most accurate way to see if the server can be reached. The first thing that 
comes to mind is that I should probably look for a specific response, otherwise 
a proxy server returning an error would also appear to be a success.

- Ian

On May 15, 2014, at 10:52 PM, Andrew W. Donoho <[email protected]> wrote:

> 
> On May 15, 2014, at 21:49 , Ian Ragsdale <[email protected]> wrote:
> 
>> Does this make sense? Does anybody see any obvious problems with this?
> 
> 
> 
> Ian,
> 
> 
> 
>       I have a somewhat intimate relationship with the Reachbility library. 
> Are you sure it ever leaves the machine? In my experience, it solely depends 
> upon local state of the radios, etc. Touching a host, the recommended way to 
> start the radios, does leave the machine. But that isn't a Reachability 
> function. The first thing that happens, of course, is name resolution. That 
> is normally where things start to fall apart with my connectivity.
> 
>       Hence, I'm not sure I'm buying your analysis of the data nor your 
> identification of the culprit class.
> 
> 
> 
> Anon,
> Andrew
> ____________________________________
> Andrew W. Donoho
> Donoho Design Group, L.L.C.
> [email protected], +1 (512) 666-7596, twitter.com/adonoho
> 
> Download Retweever here: <http://Retweever.com>
> 
> No risk, no art.
>       No art, no reward.
>               -- Seth Godin
> 
> 
> 

-- 
You received this message because you are subscribed to the Google Groups 
"Couchbase Mobile" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/mobile-couchbase/EFA1D1D7-FCC9-4E9C-AE69-5B97D5279E91%40gmail.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to