Thanks, I am happy!

On Sun, Nov 18, 2012 at 12:47 AM, Christian Grothoff <[email protected]
> wrote:

> On 11/18/2012 01:21 AM, Helin Wang wrote:
>
>> Hi Christian,
>>
>> Thanks for your response.
>> I did more tests, seems post processor can be created but won't be called
>> if I did not specify encoding. post processor can't be created if I use
>> GET. Event more, with GET (curl -d 'data' -G), *upload_data_size will
>> always be 0.
>> Anyway, as you said, parse myself should be the best solution. I am using
>> it, with POST (*upload_data_size will always be 0 with GET). So I found my
>> way, thanks:)
>>
>
> Well, parsing  the data yourself or not is clearly secondary here; by
> switching to POST you now at least do *get* some data (and now you can
> decide how to parse it).
>
>
>  Hi Christian, and everyone,
>>
>> There is another issue blocking me now.
>> I am using threading model MHD_USE_SELECT_INTERNALLY, if I sleep 3s in the
>> answer function.
>>
>
> There's already your problem: you are not "allowed" to block in the answer
> function if you use MHD_USE_SELECT_INTERNALLY as the threading model.  Then
> we have 1 thread, and you blocking it will have seriously bad consequences.
>  If you want to do blocking operations like that, you should use the
> one-thread-per-connection model.
>
> Happy hacking!
>
> Christian
>
>

Reply via email to