2

(A follow up question can be found here: link)

In an attempt to understand how the packet encoding blocks work I tried to mimic the experiment performed by Aaron, published in this post, but reduced the flowgraph to the components that interested me the most, namely the Packet Encoder, the Packed to Unpacked and the Packet Decoder. I have also used a text file in_strings.txt instead of the cat_in.png.

Follows the flowgraph I have implemented:

enter image description here

My expectation: I was hoping that out.txt (File Sink) would be equal to in_strings.txt (File Source).

What really happened: out.txt is very similar to in_strings.txt but it contains additional unwanted x characters that show up after 65535 expected characters.

Follows a preview of the diff side by side between the input and the output:

enter image description here

Follows the in_strings.txt file: https://pastebin.com/NWa7NEww

I wonder why those x's show up in the output?
Thanks in advance!

fmagno
  • 205
  • 2
  • 7

1 Answers1

3

Packet Encoder and Decoder are broken; they drop data.

That's why they are in the deprecated category (for years now!). We've removed them, because as a project, GNU Radio has not been able to fix them (and also, they were terrible from an architecture point of view).

So there's exactly one solution: don't use packet encoder / decoder.

Marcus Müller
  • 30,525
  • 4
  • 34
  • 58
  • 1
    thanks! What are the alternative blocks and which would be the equivalent flowgraph? – fmagno Mar 22 '19 at 21:48
  • 1
    Depends! I'd probably go for the architecture you find in the GNU Radio examples, typically installed to /usr/[local/]/share/gnuradio/examples/digital/packet, especially packet_rx.grc, packet_tx.grc and packet_loopback.grc – Marcus Müller Mar 23 '19 at 08:03