More maintainers of avra

[tomek-szczesny]

I admit I play with AVRs on and off every few months (years when there's a shortage) and every time I come back and rebuild a toolchain, there's a problem with outdated avra in mainstream Linux repositories that misses support for the newest AVRs.

Adding support for more AVR UCs is trivial, much of that hangs in the pull request section and I could provide the rest, and double check PRs in this matter.
Others have identified different issues and I trust some of these contributors would be willing to help maintaining avra upon receiving an invitation.

My goal here is to offer a hand with some of that and help @Ro5bert review PRs, hopefully ending up with a long overdue release.

[Ro5bert]

I have no time to or interest in maintaining AVRA right now. I recognize that you've used and contributed to AVRA a lot, so would you like me to transfer ownership of the AVRA repository to you?

[tomek-szczesny]

I'd rather avoid confusion caused by moving the repository to a different account, now that it has some 40 other forks.
I suggest adding me as a collaborator for now, so I can deal with PRs and add some modifications of my own, and roll a release.
If everything works out I'll ask you for a complete transfer.

What do you think?

[Ro5bert]

Ok, I've invited you as a collaborator. Let me know if you need anything more from me (e.g., add other collaborators), or if you want to transfer the repo.

[tomek-szczesny]

Thanks, it is done.
Just wanted to ask real quick about the dev branch - how complete are these changes, and is it worth spending time studying that?

[Ro5bert]

It's been a while, but I think the extra code in dev is related to a large release I was planning ("AVRA 2.0"). That release wasn't to add any new features, but rather clean-up code. You can find some of the large changes I was planning briefly described in the TODO file.

The main new stuff on the dev branch seems to be

• new argument parsing in newarg (not integrated) to simplify args.[ch];
• a hash map implementation in hash.c (probably not used anywhere yet) to replace the linked lists used everywhere currently, which, e.g., lead to performance quadratic in the number of labels in an input file (yikes!);
• a new lexer (not integrated) to help clean-up the mess in parser.c (I was planning on replacing the entire parser, and I think I wrote some code for that, but I'm not sure where it went).

I also recall doing some work on improving the flexibility of device capability specification (i.e., improve device.c), but I'm not sure where that went either. Perhaps that was only design work, and there was no code for it yet; I don't remember.

I leave you to decide if you want to keep any of that code. I do think AVRA has major code quality issues, hence the planned internal changes. Basically every time I looked at the code base, I found new bugs. In particular, mneumonic.c is horrific.

[tomek-szczesny]

Thanks for your comments. I guess I won't be touching that for now and focus on accepting a few PRs.

[CableGuy67]

This was curious to find when I was cleaning out old repositories.

Just to weigh in the big reason I forked was to add and have a bit of control some fixes added by @srtlg that enabled avra to be used to build AmForth and move away from the wine method that had been used. Pretty much that was the application of their PRs #52, #55 and #56. I did some simple checking and found that my patched avra gave the same output as using avrasm.exe with wine for AmForth.

I poked around a little bit to find out why the single quote was short circuiting things but didn't really pursue it. I see now that someone else had another issue since then that looks like it directly relates in #73.

I just wanted to stick my toe in the freshened water since I am one of the forks mentioned and one of the few that actually had changes after forking. I had intended to bring the avra repo that I forked into AmForth as maybe a submodule (or just properly patch in the needed files to preserve history) but never bothered since it Just Worked for what I was doing at the time and only really use it for an already supported ATmega1284p.

In any case, it's always nice to see activity.

[tomek-szczesny]

Hi @CableGuy67 ,

Sorry it took me a while but I'm looking into all this today.
So if I understand correctly your independent version of AVRA would be a nice addition to this mainline repo, and includes a few existing PRs, with added tests and doc corrections. I can't see anything that could potentially break the preexisting functionality even though you said it's a version dedicated for use with some other project.
Frankly everything you've done except the very last commit (README modification) would be welcome as a PR. Do you think you could do it? I'm not a git whiz and would rather accept these changes as a PR instead of trying to pull them myself. :)

Thanks for chiming into this conversation. My goal here is to collect all those improvements made by the community in one place.

[CableGuy67]

I believe that everything I did that you'd be interested in is here as a PR already. I'll look into this in the next day or so since I'm on summer holiday. I'd have to verify what exactly is going on and see if there is anything extra I did to any of those. IIRC though, it was mostly @srtlg 's work I applied to @Ro5bert work.

I will double check though to be sure.

[CableGuy67]

Looking back it seems that I started from commit afdbda4 and then applied:

PRs: #52, #55 and #56. The first and last being the most important one to get things working for AmForth. I also applied the changes authored by the same user of commit 614f914 from their development branch so that one could be cherry picked in or whatever you feel is best.

None of the things applied were my own work. I merely ferreted out changes that had been sitting around and put them together to get things working so that everything could be built without having to resort to the windows exe.

Grabbing all those will be sure to keep the commit history intact without the added noise of interim fork. Let me know if there is anything else I can clarify for you. I had to rattle my brain cells to remember just what I had done with all this. :)

---EDIT---
Now that I look again it's even simpler. All I really did was to merge their development branch into master. That makes it easier to see. They just compiled the 3 PRs that they had made here and added in the typo fixes to constitute that branch.

[saper]

Wow, coming back here in 2026 ... to compile AmForth :) thanks for keeping this alive. Let's just merge this branch and move on...