I tried to decode the ir_nec protocol but no decoding information appears on the screen. I also tried with another IR protocol, and there was some decoding info, although wrong since it was for a different protocol. So I am rather sure there is a ir_nec protocol specific issue. I tried at two protocol stages: 1. demodulated signal after a dedicated receiver 2. modulated signal directly on the transmitting IR diode (not sure I really tried this) The issue should probably be apparent with a recorded sample, but I did not check yet. I used a cheap FX2 hardware, but this should probably not matter. The Windowas 32bit PulseView nightly from 5 of March 2018 was used.
I checked again, and the protocol decoder works fine if a modulated signal is provided, but not at all with a demodulated signal (after a common IR receiver). So instead of a bug report this should be a feature request. There is a small problem with the decoder configuration UI, there is no unit specified for the modulation frequency. The unit is [Hz] which is easy to guess, but not obvious, since it is usually specified in [kHz]. Also the current default value is 0, 48000 (48kHz) would be a better choice, since this is the frequency I encountered in all documents regarding this specific IR NEC protocol.
I now downloaded some signal dumps: https://sigrok.org/gitweb/?p=sigrok-dumps.git;a=tree;f=ir/nec/hama_8in1/tv_matsui_0001;h=6a7407c34054449832ee22e4d3cadb91596db981;hb=HEAD I downloaded the latest nightly and had no problem decoding the dumped signals. So I am not sure what could be the issue. But while reading the source code I noticed a constant which is probably wrong: https://sigrok.org/gitweb/?p=libsigrokdecode.git;a=blob;f=decoders/ir_nec/pd.py;h=02d70a9c299aa36ce5fffcb6fc41075715cfeb9c;hb=HEAD#l123 The constant should be 0.5625ms and not 0.652ms.
Can you provide a specific *.sr file for sigrok-dumps (taken using your fx2lafw LA) where you see the issue, so we can reproduce? Thanks!
Closing the report. The stop interval got adjusted in 2020-08, no dump has become available for the issue reported here. There has not been an update in a long time.