unpredictable crashes using http client

[extrawurst]

i am really lost here. since updating from stable 0.7.19 to the current alpha i get random crashes in my http client app:

(normally i am used to getting nice callstacks and all)

[02540F80:02544500 trc] ---------------------
[02540F80:02544500 trc] HTTP client response:
[02540F80:02544500 trc] ---------------------
[02540F80:02544500 trc] HTTP/1.1 200 OK
[02540F80:02544500 trc] Server: vibe.d/0.7.20
[02540F80:02544500 trc] Date: Tue, 02 Sep 2014 18:53:00 GMT
[02540F80:02544500 trc] Keep-Alive: timeout=10
[02540F80:02544500 trc] Access-Control-Allow-Origin: http://localhost:8080
[02540F80:02544500 trc] Access-Control-Allow-Credentials: true
[02540F80:02544500 trc] Connection: keep-alive
[02540F80:02544500 trc] Cache-Control: no-store, no-cache, must-revalidate, max-age=0
[02540F80:02544500 trc] Content-Type: text/plain; charset=UTF-8
[02540F80:02544500 trc] Content-Length: 498
[02540F80:02544500 trc] ---------------------
[02540F80:02544500 trc] evbuffer_read 498 bytes (fd 644)
[02540F80:02544500 trc]  .. got 498 bytes
[02540F80:02544500 trc] read data
[02540F80:02544500 trc] exception trap
[02540F80:02544500 ERR] Task terminated with unhandled exception: Win32 Exception

any ideas ?

shouldn't the next line of code logDebug("Full error: %s", th.toString().sanitize); write out more info ?

[etcimon]

Took me a while to fix my pull request, but the latest commit should be good and should fix this error. The problem is a bad keep-alive value at 0 seconds, it closes the connection unexpectedly and causes subsequent read/writes to fail.

#756

[extrawurst]

Ok, well i am happy to test it once it is tagged ;)

[s-ludwig]

#756 got merged, please retest with 0.7.21-rc.1 once you get some time.

[extrawurst]

for me this issue seems to be fixed, i am closing this one!