Crash after second request to not accessible target about asynchttprequest HOT 9 CLOSED

from asynchttprequest.

Here you go... I've left some of my debug messages in there which you (hopefully) will be able to match to the scenario above.

Note that the "crash" occurs right at the time that the response from the third request (the one launched with a http.readyState() of 0) would be expected to arrive.

[6709] Ready State: 0
[6709] Push Sent
Debug(6710): setDebug(on) version 1.1.15
Debug(6710): open(GET, http://192.168.59.184:2002/?valu)
Debug( 1): _parseURL() HTTP://192.168.59.184:2002/?value=1
Debug( 6): _connect()
Debug( 11): send()
Debug( 11): _buildRequest()
Debug( 12): _send() 126
Debug( 14): *can't send
Debug(30127): _onError handler error=-13
Debug(30128): _onDisconnect handler
Debug(30128): _setReadyState(4)
[36838] HTTP Status Code:-4
Debug(30129): responseText() Debug(30131): responseText() no data
[43161] Ready State: 4
[43161] Push Sent
Debug(36452): setTimeout(2)
Debug(36452): setDebug(on) version 1.1.15
Debug(36452): open(GET, http://192.168.59.184:2002/?valu)
Debug( 1): _parseURL() HTTP://192.168.59.184:2002/?value=0
Debug( 5): _connect()
Debug( 9): send()
Debug( 10): _buildRequest()
Debug( 12): _send() 126
Debug( 14): *can't send
Debug(30845): _onError handler error=-13
Debug(30845): _onDisconnect handler
Debug(30846): _setReadyState(4)
[74011] HTTP Status Code:-4
Debug(30846): responseText() Debug(30849): responseText() no data
[80021] Ready State: 4
[80021] Push Sent
Debug(36857): setTimeout(2)
Debug(36857): setDebug(on) version 1.1.15
Debug(36857): open(GET, http://192.168.59.184:2002/?valu)
Debug( 1): _parseURL() HTTP://192.168.59.184:2002/?value=0
Debug( 5): _connect()
Debug( 9): send()
Debug( 10): _buildRequest()
Debug( 12): _send() 126
Debug( 14): *can't send
[88459] Ready State: 0
[88460] Push Sent
Debug(8435): setTimeout(2)
Debug(8435): setDebug(on) version 1.1.15
Debug(8435): open(GET, http://192.168.59.184:2002/?valu)
Debug( 1): _parseURL() HTTP://192.168.59.184:2002/?value=1
Debug( 6): _connect()
Debug( 9): send()
Debug( 9): _buildRequest()
Debug( 12): _send() 126
Debug( 14): *can't send
Debug(21244): _onError handler error=-13
Debug(21245): _onDisconnect handler
Debug(21245): _setReadyState(4)
[109709] HTTP Status Code:-4
Debug(21246): responseText() Debug(21249): responseText() no data

--------------- CUT HERE FOR EXCEPTION DECODER ---------------

from asynchttprequest.

I had a quick look at this. The restart appears to be caused by (in my example above) the second connection attempt de-allocating _client when the onDisconnect handler fires while the the third connection attempt is still active. The third attempt then tries to re-use the already deallocated _client causing the restart. As I said, I can work around this by looking for the http.readyState() to be 4 (not 0) (after the first connect attempt) and queuing/retrying subsequent messages until the previous attempt has timed out. This stops the resets.

However, we still have the issue that there is a (+/-) 30 sec timeout before the onDisconnect event turns up even though the application "knows" it "*can't send" after less than 20 ms. Any thoughts on how to abort the connection and/or fire the onDisconnect event? I've tried using the abort() command after a timeout however that has no effect. Any thoughts appreciated.

from asynchttprequest.

Hello, I have the excact same problems.

Here is my example code and here is the reboot msg.

from asynchttprequest.

@andrewk123 did you solve the timout problem?

from asynchttprequest.

I tried AsyncHTTPRequest_Generic and found the same issue, however Khoi was able to resolve it so I use that version now. It's functionally identical to asyncHTTPrequest. There were some small changes to integrate it but nothing that took more than a few minutes. The fix was in 1.4.0.

https://github.com/khoih-prog/AsyncHTTPRequest_Generic

from asynchttprequest.

@andrewk123 Hi first of all thanks for your answer!

Thats actually pretty funny because I struggeled for weeks to get either boblemaire's or khoih-progs library to work properly.
After days of no success I created a issue which let khoih to make a new version with a fix. Here is his answer to my issue.

Unfortunately till today I didn't get a working version..
After the "fix" it still takes always 18 seconds to get a bad connection response. You can see my demo-code + debug output here.

Which microcontroller do you use?

from asynchttprequest.

Looks like a solution was found in another's fork.

from asynchttprequest.

@boblemaire Wich one?

from asynchttprequest.