I'm working on step 3 right now, which is going to be a lot bigger than step 2. Step 3 gets rid of the rest of the .hpp's and the rest of the "loose" files in the source tree.

Yes, there was a hodge-podge of .h and .hpp files. I've always preferred .h even for C++ because you tend to look at the end of the file (the extension) to pick out the headers from the source files, and it's more immediately clear if .h is used.

I see no reason to separate source files into a src folder, and headers into an include folder for this project. It's not that big. Moving headers to an include folder involves more paths. If a function is added, it's declared in the header anyway, and doing it this way keeps the two together.

I pushed an intermediate commit for step 3. Working on step 4. I'm doing this manually with Xcode, and running continuous code validation to have Xcode point out what's broken so it can be fixed as I go.

Step 4 should finish up the base tree.

Step 4 makes it look like all the code disappeared. Builds and runs fine.

@wmiler I did not move those .desktop files, did not want to break your workflow. They can be moved to JS8_Include when you get time up update your workflow path(s).

@wmiler I did not move those .desktop files, did not want to break your workflow. They can be moved to JS8_Include when you get time up update your workflow path(s).

What's message_aggregator.desktop? That's not used in any of the workflows to my knowledge.
Looks like it was last touched 9 years ago, is this something from the original creation @jsherer ?

Also, affected includes have been resorted with LLVM/clang-format in Xcode. The order of some of the includes was pretty bad, kinda like stuff just kept getting added on in random fashion. So I fixed that.

I'm not sure what message_aggregator.desktop is. It doesn't appear to be used anywhere. So if it is a dead file, or unless somebody knows where it came from, it can be removed.

Xcode commit history shows message_aggregator.desktop appeared in the source tree back in 2018 with Jordan's initial commit and hasn't been changed since. So it probably came from wsjt-x and it's a dead file.

As it is, merging this will break two outstanding PR's. Which would not be fair to those developers to force them to rebase and fix conflicts. So I changed the label from ready to do not merge until those are dealt with.

This is complete. Processes done:

• Code cleanup - move loose source and header files in source tree to group folders
• Rename all header files with .h extension, previously a random mix of .hpp and .h header files
• Fix function declaration in HamlibTransceiver.h that overrode member function but was not declared override
• Remove unused variable in Modulator.cpp
• Remove unused files from source tree
• Remove duplicate LazyFillComboBox source and header files
• Reformat codebase to LLVM C++ standard

I'm happy with it. Would recommend merging ASAP as if there is any outstanding work this will seriously break it. It was non-trivial to fix the conflicts with the WSJTX UDP protocol commit, then move that commit into the new source tree.

I'm ok with landing this now, rather than later.

Merged, has the required number of reviews and passes all the CI checks. The codebase should now be much prettier and hopefully easier to understand.

One further comment on this. There is a JS8_Deprecated group that has classes marked for deprecation. I did not move OmniRig into it at this time. Even though OmniRig is technically deprecated there still exists the possibility that somebody might revive it, as it was a popular Windows program at one point.

So I left OmniRig in the JS_Transceiver group for now. I figure we'll give that some time and if nothing ever happens with it we'll move it to the JS8_Deprecated group, and from there it has about 6 months and it goes away.