Thanks for answering. 

Indeed I found a pattern and probably the explanation, now I need to found the 
appropriate solution:
What happen is that My iphone application presents the data that comes as an 
xml files, which I fetch from a various urls on the same server. There are 
problematic urls, that when I try to reach them from safari or from within the 
application give an error - mostly timeout error. When I initialize the 
NSURLConnection I define a timeout in my NSURLMutableRequest in order not to 
stuck the GUI and let the user to work with application. On timeout I get 
didFailWithError in NSURLConnectionDelegate and show a message that there was 
not response from the server and the user can continue working. The matter is 
that the connection or underlying socket somehow is preserved and the 
connection with the server is not closed. So when I try to get another url from 
the same server I can't reach it since the server is still stuck with the 
previous problematic request! 
Now I can understand why from the other device or the browser  I can reach 
urls, that are not reachable after problematic request on my device, simply 
because they open another socket to the same server. 
The question is how do I totally close (force closing) the problematic 
connection, that is stuck till the server will "get out" of this unsuccessful 
fetch? I tried close connection and cancel connection, but seems these APIs do 
nothing :(. Is the NSURLConnection way to go in such cases? Each time I need to 
get xml file I create another NSURLConnection. I read somewhere that it's not a 
good practice...

Again, thanks for the help. 

On May 29, 2010, at 5:04 AM, Jerry Krinock wrote:

> 
> On 2010 May 28, at 02:58, Nava Carmon wrote:
> 
>> When I enter same link in Safari it takes a second to bring data.
> 
> I understand that you're saying it "works OK at the same time from the same 
> IP address from Safari", meaning that you have ruled out being throttled by 
> this "certain server", or your ISP.  OK, because these would be the first two 
> suspects on my list.
> 
>> And after another trial I can access the link.
> 
> I'm assuming you mean exactly the same URL.  Some servers give better service 
> to web browsers than to apps requesting via their API.
> 
>> What might be the reason for such a weird behavior?  How can I improve it?
> 
> You're going to have to discover a pattern.  For another data point with 
> these types of problems, you'll usually want to open a Terminal window and 
> send your request using the unix command curl(1).
> 
>> What is the role of cache policy with NSURLConnection?
> 
> Long answer:
> 
> http://developer.apple.com/mac/library/documentation/cocoa/conceptual/URLLoadingSystem/Concepts/CachePolicies.html
> 
> For consistency in testing, you'll probably want to use 
> NSURLRequestReloadIgnoringCacheData.
> 
> Let us know if you find anything interesting.  Actually, a better list for 
> this question would be [email protected].
> 
> _______________________________________________
> 
> Cocoa-dev mailing list ([email protected])
> 
> Please do not post admin requests or moderator comments to the list.
> Contact the moderators at cocoa-dev-admins(at)lists.apple.com
> 
> Help/Unsubscribe/Update your Subscription:
> http://lists.apple.com/mailman/options/cocoa-dev/ncarmon%40mac.com
> 
> This email sent to [email protected]


Nava Carmon
[email protected]

"Think good and it will be good!"

_______________________________________________

Cocoa-dev mailing list ([email protected])

Please do not post admin requests or moderator comments to the list.
Contact the moderators at cocoa-dev-admins(at)lists.apple.com

Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com

This email sent to [email protected]

Reply via email to