Block2 request does not seem to notify CoapHandler#onError if exchange stucks on non-first request. about californium HOT 14 CLOSED

Hi @evgeniynickolaev,

this is a very good observation you have made. Thanks for pointing this out and doing the debugging :-)
Your proposal for fixing it goes in the right direction, I think. However, we will need to distinguish between requests originating locally vs. requests we are sending a response to.

Would you like to create a PR for fixing it? If so, make sure to read the contribution guide ...

from californium.

Thanks for reply!
Yes, I am very keen to contribute, I'll work on it.

from californium.

Cool :-) Looking forward to your PR (and discussing it) ...

from californium.

This is the PR.
Now it invokes CoapHandler#onError.

Not sure if checking the message class is sufficient to cover:

However, we will need to distinguish between requests originating locally vs. requests we are sending a response to.

Also I've found this case too special to be included in BlockwiseTransferTest.

from californium.

However, it would still be good to have a test case for this.
I think you can create one along the lines of the org.eclipse.californium.core.test.lockstep.BlockwiseClientSideTest ...

from californium.

OK, will add

from californium.

I've pushed a test case. I don't really like it but wanted to confirm if the approach is appropriate at first.

from californium.

I think the test itself does exactly what it should do 👍
Why don't you like it? ;-)
I have added some proposals for simplification ...

from californium.

Oh, BTW: please rebase your branch onto current master ...

from californium.

Thank you)
Is it OK to smash my commits into one during rebase and push it with force?

2016-04-22 10:40 GMT+03:00 Kai Hudalla [email protected]:

Oh, BTW: please rebase your branch onto current master ...

—
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub
#36 (comment)

from californium.

Absolutely

from californium.

Done

from californium.

Hi, also thanks from my side on spotting and fixing this.

I noticed that we are more and more smeering the layers. While this is a valid fix for now, I will try to solve it purely from the BlockwiseLayer, probably by attaching a MessageObserver there on the intermediary blocks.

from californium.

Good idea, @mkovatsc 👍

from californium.