On 4/7/15, 2:38 PM, "Thiago Macieira" <thiago.macieira at intel.com> wrote:
>On Tuesday 07 April 2015 11:24:05 Kesavan, Vijay S wrote:
>> There are sample applications to showcase the usage of the C-SDK at
>> resource/csdk/stack/samples
>
>Ah, I see. There's a reason I didn't find them: they're written in C++. I
>was
>looking for C sources...
>
>Anyway, I had the qualification of "good" examples. I don't consider this
>a
>good example:
>
>static std::string putPayload = "{\"state\":\"on\",\"power\":5}";
>
>The payload is hardcoded! An example application should be teaching how
>to do
>things right...
>
>This sample application is also designed to be a testcase, as seen by a
>variable (!) called TEST_CASE. That has implications on what the
>application
>does and doesn't do, since it's not designed to be a teaching tool.
>
>Another part I feel quite strongly about is this:
>
> while (!gQuitFlag) {
>
> if (OCProcess() != OC_STACK_OK) {
> OC_LOG(ERROR, TAG, "OCStack process error");
> return 0;
> }
>
> sleep(2);
> }
>
>No example, under any circumstances, should EVER use the "sleep"
>function.
>There are very few valid uses for the sleep/msleep/usleep/nanosleep/etc
>family
>of functions and most novice developers will not know them[*]. We need to
>write our examples in the way that real applications should be using the
>stack
>and clearly this isn't it.
>
>[*] given that this is in our code, it's clear also that many advanced
>developers don't know either.
>--
In the C++ SDK (InProcClientWrapper , InProcServerWrapper)::listeningFunc
has the same pattern.
My understanding (before the CA branch) was that it was because the
OCProcess() wasn?t using any mutex?s/conditional variable
So the sleep was a hack to prevent a pegged thread.
Looking in the bowls os OCProcess, it looks like it is now using a mutex,
but the sleeps are still there.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5859 bytes
Desc: not available
URL:
<http://lists.iotivity.org/pipermail/iotivity-dev/attachments/20150407/56446d49/attachment.p7s>
