Wrong work with RR, RNR, I frames

By trgv , 4 August, 2023
Forums
Gurux.DLMS

As we get RR frame with NS and NR = 0
Subtest 2: Check that the IUT is in NRM
Send an RR frame. Expect an RR frame.
This test went wrong...

Using C++ lib

Kurumi

2 years 9 months ago

Hi,
Can you tell me a little bit more? Are you talking from GXDLMSDirector T_HDLC_FRAME_P2 test?

BR,
Mikko

trgv

2 years 9 months ago

In general, the library does not handle frames of this type correctly. For example, when receiving a frame with NS and NR == 0, the client or server must respond with the same frame and not increment the frame number.

Kurumi

2 years 9 months ago

Hi,

ANSI C will pass the tests because it's usually used with microcontrollers and that is the language that our clients are using.
ANSI C++ is used usually in client-side applications. I need to think should ANSI C++ also pass the DLMS tests.

BR,
Mikko

trgv

2 years 9 months ago

Since our controller supports embedded linux, it was decided to use the C++ library. Here we are faced with the above problem, one of them.
Take a closer look at my PR on Github

trgv

2 years 9 months ago

In your code, when receiving an RR frame with N(R), the frame increment occurs, but it should not. RR should be returned with N(R) equal to the current receiver frame. If previous frame we send wasn't I frame.