|
|
Screencast From 2025-08-18 11-43-02_ ardourbug.webm (768,275 bytes) |
|
|
also, with sections of automation like this. there should be a node at the edges of the automation late for each clip. here it just randomly goes off the side, where i cant grab or edit it. |
|
|
any update? this bug was super annoying to work around. |
|
|
ok, this is getting ridiculous lmao. again, i asked ai to fix this issue and it did it in less than a minute. its a single line fix. at this point i feel like my bug reports are just being ignored. this was super easy to fix. anyways hope this note helps. image.png (49,612 bytes) |
|
|
@pyrotek45 I myself cannot confirm whether or not your three AI-given fixes are solid (i.e. won't interfere with other code elsewhere), but, if they are, it does beg the question: Will Paul or Robin ever utilize the help of AI recommendations for easier (but numerous) bug fixes/requests? Based on their previous sentiments, I would guess not. Now, in general, they do currently seem to be sidelining about 90% of incoming bug reports, unfortunately. I think mostly because they’re working on bigger things (-like restructuring of the pianoroll, etc.). But you have to remember that they’re basically a *two-person* team, which is incredibly tiny for such a beast of a project and inflow of bugs/requests, -regardless of AI-use or not. … Also, I know from a lot of experience using AI to help with coding, the results given look SUPER convincing ***on the surface***. But, if you do it for long enough, you will discover that the AI (-even one specifically ‘trained’ for your exact software/code/DAW/etc.) is always incredibly *nuclear* in its approach, generally lacking broader awareness, which ends-up making critical errors over and over again with respect to *the larger context* of the codebase/goals. So, all in all, it’s wise to approach any AI ‘solutions’ with *a lot* of caution, instead of ‘oh, it’s totally fixed. all is good! wow, so easy!’, as I did for a long while, until I learned better. : 0 : ) Thanks for reading, -J |
|
|
@ghostonacid here are my thoughts. after a year of waiting, i figured using ai to probe the code and see if it could fix the issue was a good idea. so i did. after adding the fix, and recompiling ardour ( which takes a long time btw ), the fix worked. as with all the other fixes ive asked ai to do. sure there might be some unforseen effects ( even though the fixes are but a handful of lines of codes to add, a single character change in this case and something which is true of human code/ bug fixes too), but i suspect paul or x42 would be able to see if the fix is good or not. whether or not they accept the fix ( since i did eventually make them pull request on github ), i figure, they will at least be more aware of where to go or how to fix the issue on their end. regardless i only see this is a means to get ardour to be in a less buggy state. my main goal with this, is to help them do some of the leg work to fix the bugs, by any means necessary. esp after waiting years and seeing many of my reports go unnoticed. even though many of them seem to be fixed with just a few lines of code, it does make me wonder if my reports ever even got a second or two looked at. |
|
|
@pyrotek45 For sure. I feel your sentiment 100%. And I also didn't realize your report(s) was as old as it was... O___o Yes, most likely for changes that require such tiny intervention to fix, AI solutions are *probably* just fine. : ) I just wish their dev team was bigger... : ( Because yes, there seems to be endless bug reports on here, and also endless opportunities for just *polishing* Ardour. (-Which I can't stress enough. O____O) I imagine working on NEW features is more alluring to them then polishing/correcting/fine-tuning that which already exists. But polishing what you already have is *absolutely vital* to making your software just, well, 'feel right'. Again, they could greatly benefit from *at least* one more full-time developer. I hope that happens one day... But who knows? -J |
- Anonymous
