Re: [libmicrohttpd] Correct handling of the MHD_ContentReaderCallback

[Christian Grothoff]

Yes, I can confirm that this should not happen by design.

Happy hacking!

Christian

On 02/10/2014 11:48 AM, Martin Velek wrote:
> This is only "what if" question. I am using the MDH on an embedded
> device and I need to take into consideration all possible input
> alternatives. I am using two assertions, NULL != buf, max > 0, they
> were never activated.
> 
> If you can provide an assurance that it should not happen, it is not
> an intented usecase, that's ok for me (mental peace). I did inspection
> of the source code and I think it should not happen.
> 
> Best
> Martin
> 
> On Mon, Feb 10, 2014 at 11:19 AM, Christian Grothoff <address@hidden> wrote:
>> Hmm. That's, eh, interesting.  MHD should simply not call you
>> asking for zero bytes.  Got a testcase or a (gdb) call trace
>> for me? Or is this a hypothetical question to answer the case
>> "what if MHD did do that?"?
>>
>> Happy hacking!
>>
>> Christian
>>
>> On 02/10/2014 10:50 AM, Martin Velek wrote:
>>> Hello,
>>>
>>> I am trying to write a robust MHD_ContentReaderCallback callback.
>>> However I cannot imagine how to handle a situation when MHD calls my
>>> callback with the provided buffer size equal to 0 (I am running
>>> internal select mode). The documentation does not state it as an
>>> impossible situation.
>>>
>>> I can't return 0 because of the internal select mode.
>>> I can't return MHD_CONTENT_READER_END_* because I would like to handle
>>> the request.
>>>
>>>
>>> Best
>>> Martin
>>>
>

0x48426C7E.asc
Description: application/pgp-keys

signature.asc
Description: OpenPGP digital signature