Daniel,

I have inserted debug displays that show that all array files have processed successfully prior to displays dealing with other processing and then displays showing the new array files starting and then abending [1 ~ 10 with a setting of 10 at a time] so while on the surface your solution doesn't appear [to me] to apply, it certainly makes a lot of sense especially since I've been fooled before with display sequencing.  Because I've used Curl in upwards of a dozen applications I've found one I wrote some while ago using an approach I have yet to try in my current effort.  I'm in the middle of coding that approach and if testing again fails, I'll most certainly try your approach.  Thanks.

Phil Welch

On Friday, February 21, 2014 5:00 PM, Daniel Stenberg <daniel_at_haxx.se> wrote:

On Mon, 17 Feb 2014, Ruslan Gaifulin wrote:
>
>> it can't even get to curl_multi_info_read out of the loop (still_running is
>> always non-zero)
>
>You should do curl_multi_info_read() for each handle that finishes. If no
>handle is finished, you know that the transfers are still going.
>
>--
>
>  / daniel.haxx.se
>-------------------------------------------------------------------
>List admin: http://cool.haxx.se/list/listinfo/curl-library
>Etiquette:  http://curl.haxx.se/mail/etiquette.html
>
>
>  

-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-library
Etiquette: http://curl.haxx.se/mail/etiquette.html
Received on 2014-02-21