Without going and posting in the Bugzilla comments as it is not my intent to cause any sort of brigading here, I’d like to ask people here if I (the bug reporter) am being unreasonable for thinking it’s absurd to close this memory corruption bug as a WONTFIX: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124226
The reason I’m asking here is because this is the 7th bug I’ve had closed as a WONTFIX. Some of the others have been lower impact, but this is the second closed memory corruption bug and I want to know if I’m really being unreasonable here by getting some opinions from third parties before I argue too much for this and other WONTFIX bugs not being closed.
I’m not saying that I think these should be fixed right now to be clear, AdaCore certainly doesn’t owe me or anyone else a fix if they’re not paying for commercial support, but closing these bugs both hides them from GCC developers who may want to patch them and hides them from users who wish to ensure that their code does not contain any code which will trigger known defects.
Edit: This was posted around the time of comment 3 on Bugzilla.
I looked at your bugreport and I think WONTFIX is the wrong approach. I agree it could lead to a waste of time debugging when you have two virtually the same code, just different objects, leading to different generated code.
Reading over this, as well as some of the other bug reports you posted(e.g. 1 & 2) , it looks like Eric Botcazou is taking the bug reports in bad faith and assuming LLMs.
No human being would use Dimension_System for a GUI!
Please stop fuzzing the compiler! No human being writes iterated associations nested in multiple layers of generic like this, that's totally unreadable.
I obviously no longer believe what you're saying. My hunch is that you're using an LLM to generate artificial Ada programs and throwing them at the compiler to see what happens.
I did notice that he fixed the bug the very same day you posted your entire code, but it doesn’t look like there’s plans to fix all of the other bugs that you submitted.
I’ve never had an experience like this when submitting a bug to GCC. I’ve always went out of my way to try and narrow down my bugs and submit the least-amount of code that produces the example instead of submitting my actual code, which seems to be what you’ve done too.
I spent a few hours going through all of my previous bug reports and showing the exact code where they came from, as much as that was possible, here. Two of the other comments above came after that and many of my reports still remain closed or have been closed since then.
I really would prefer this to not be a thing that creates a whole lot of drama, but I don’t see any way that this can be resolved without me posting more about it publicly. I just want to be able to open bug reports which can be fixed one day by people who know how GNAT works better than I do so I can eventually use those features in my software.
Again I am not asking for bugs to be fixed right now or on any sort of timeline. I just want them to remain open so others are aware of them and can choose to donate their time to fix them. I am completely fine with them remaining ignored for an indefinite amount of time. I have said as much to Eric here.
When they’re all just being closed no one will ever see them or fix them. In some cases I have spent a significant amount of time working on patches (1, 2, 3), and it is obviously very frustrating to have that work thrown out by a maintainer that has a grudge against me. I didn’t think too much of the first two but in the third I was again accused of fuzzing the compiler after spending the time to provide evidence that I am not, so this seems to be some kind of issue that he has with me personally.
I also want to note that Eric has spent a lot of time fixing my other bug reports prior to this and fixed many bugs that I could not have without spending many years studying the GNAT internals. I would assume he’s spent upwards of a hundred hours working on bug reports submitted by me. This is not just a case of him coming in and taking issue with me out of nowhere. I don’t want people to miss that fact if they do not go look at the full list of my bug reports themselves.
I have a feeling that the increase in AI agents submitting things to open source content is causing maintainers to treat more submissions in bad faith.
I don’t know what recourse there is here; I don’t know who works on GCC that’s on this forum, maybe @sttaft has some thoughts.