level_pedestals now has its daq format set by DAQ->SETTINGS->FORMAT

[joshgreaves332]

added tasks/level_pedestals.h to algorithm/level_pedestals.cxx and replaced hardcoded SIMPLEROC with pftool::state.daq_format_mode, set by DAQ->SETTINGS->FORMAT (line 50 ish)

[tomeichlersmith]

So, in order to get this to compile, you will need to update the stuff in algorithm and tasks that use the DecodeAnd* classes to create different ones depending on which type of format is being used. I did this as an example in daq.cxx:

pflib/app/tool/daq.cxx

Lines 352 to 380 in 29931a0

	if (decoding) {
	std::unique_ptr<DaqRunConsumer> consumer;
	switch(pftool::state.daq_format_mode) {
	case Target::DaqFormat::ECOND_SW_HEADERS:
	consumer = std::make_unique<DecodeAndWriteToCSV<pflib::packing::SoftWrappedECONDEventPacket>>(
	fname+".csv",
	[](std::ofstream& f) {
	f << std::boolalpha;
	f << pflib::packing::SoftWrappedECONDEventPacket::to_csv_header << '\n';
	},
	[](std::ofstream& f, const pflib::packing::SoftWrappedECONDEventPacket& ep) {
	ep.to_csv(f);
	});
	break;
	case Target::DaqFormat::SIMPLEROC:
	consumer = std::make_unique<DecodeAndWriteToCSV<pflib::packing::SingleROCEventPacket>>(
	fname+".csv",
	[](std::ofstream& f) {
	f << std::boolalpha;
	f << pflib::packing::SingleROCEventPacket::to_csv_header << '\n';
	},
	[](std::ofstream& f, const pflib::packing::SingleROCEventPacket& ep) {
	ep.to_csv(f);
	});
	break;
	default:
	PFEXCEPTION_RAISE("BadConf", "Unable to do live decoding for the currently configured format.");
	}
	daq_run(pft, cmd, *consumer, nevents, pftool::state.daq_rate);

In order to avoid all this boilerplate, you can maybe use a helper function that implements the switch and exception for us. This helper function could live in daq_run.h (and daq_run.cxx).

[tomeichlersmith]

⚠️ warning I am going to drop SoftWrapped* and just update MultiSample*, but I am going to wait until after the firmware changes I chatted with Jeremy about yesterday.

#246 (comment)

TLDR: we're going to use the SoftWrapped* header, but add a special trailer word to mark the end of a MultiSample*.

These changes will make this branch and its base branch 246-econd-live-decoding kinda obsolete. When I make these updates, I will work on updating the "live" decoding in the same way that I was planning with 246-econd-live-decoding but with MultiSample* and not SoftWrapped*.

[cmantill]

@tomeichlersmith do you prefer in that case that we close this branch and wait for your changes to make changes to level_pedestals?

[tomeichlersmith]

We definitely can, I just wasn't sure how much Josh had worked on it since I reviewed it yesterday.

Edit: While closing the PR would not delete the branch, I did not want to give the impression that the work should be discarded. It can still be helpful after the refactor I mention above.