Search found 238 matches

by jarl
Fri Apr 19, 2013 8:38 am
Forum: Support - RCF
Topic: Cancel long-term RCF functions.
Replies: 1
Views: 4890

Re: Cancel long-term RCF functions.

From the client side, you can cancel the call in two ways: 1) If you are doing a synchronous RPC, you can put in a progress callback, and when the client wants to cancel, throw an exception from the progress callback. This will disconnect the client and the exception will be passed on to your own ex...
by jarl
Tue Mar 26, 2013 11:41 pm
Forum: Support - RCF
Topic: Session bindings using RCF 2.0
Replies: 2
Views: 5168

Re: Session bindings using RCF 2.0

That is correct Spino. SessionObjectFactoryService is deprecated in 2.0, and will eventually be dropped in future versions. In RCF 2.0, you should use RcfSession::createSessionObject<>() etc, to keep track of per-connection state. See this section in the user guide for more information: http://www.d...
by jarl
Tue Mar 26, 2013 10:30 am
Forum: Support - RCF
Topic: Configuration macro mismatch leads to ungracious termination
Replies: 3
Views: 6222

Re: Configuration macro mismatch leads to ungracious termina

Due to how the RCF_USE_BOOST_SERIALIZATION define is used, it's difficult to detect a mismatch at runtime. However, I've added code to the RCF initialization functions to trigger a linker error if there there is a mismatch, and that should take care of things. We should have a new build out for you,...
by jarl
Mon Mar 25, 2013 6:12 am
Forum: Support - RCF
Topic: RCF and boost.test
Replies: 2
Views: 5747

Re: RCF and boost.test

I'm not sure what's going on here. The boost.test code is almost impenetrable. The problem is that when the test fails, it is trying to print out the failed value, thus evaluating the arguments again and causing a second (failing) remote call. The client.Echo() call returns a special proxy object (R...
by jarl
Mon Mar 25, 2013 4:39 am
Forum: Support - RCF
Topic: Configuration macro mismatch leads to ungracious termination
Replies: 3
Views: 6222

Re: Configuration macro mismatch leads to ungracious termina

First off, I'm leaning towards just throwing an exception, instead of dereferencing the NULL pointer... Do you have a stack trace of the null pointer?

Triggering a linker error to detect mismatches at build-time, is definitely a good idea too - I'll have a look at what we can do there.
by jarl
Tue Mar 19, 2013 2:37 am
Forum: Support - RCF
Topic: Possible memory leak on thread pool
Replies: 1
Views: 4827

Re: Possible memory leak on thread pool

I suspect you are seeing the buffer caching that RCF does in the background. RCF reuses network send and receive buffers, rather than asking the OS to allocate new ones. To disable the caching, call ObjectPool::setBufferCountLimit() before starting your server: #include <RCF/ObjectPool.hpp> // ... R...
by jarl
Tue Mar 05, 2013 11:55 pm
Forum: Support - RCF
Topic: Boost 1.53 smart pointer changes
Replies: 3
Views: 7356

Re: Boost 1.53 smart pointer changes

Meanwhile, you can just change 'shared_static_cast' to 'static_pointer_cast' ... Boost 1.53.0 has deprecated shared_static_cast.
by jarl
Sun Mar 03, 2013 10:47 pm
Forum: Support - RCF
Topic: Boost 1.53 smart pointer changes
Replies: 3
Views: 7356

Re: Boost 1.53 smart pointer changes

Thanks for reporting this. We'll be putting out a new build shortly, to fix it.
by jarl
Mon Nov 05, 2012 3:33 am
Forum: Support - RCF
Topic: Question about client progress callback
Replies: 5
Views: 8476

Re: Question about client progress callback

Yep, I thought about suggesting that - glad you got it working.