After all the improvements I still feel like this one should be nice:
If it is not easy to find the error in your algo, what I would like to do is open a new playground, and by hand imitate the opponent until the crash occurs. However, the current hastle is that I have to manually reupload the crashing algo to pass the blockade on the playground, and then quickly open playground before it is removed again.
I’m not sure whether this is an easy fix but help would be appreciated.
Indeed, this would be nice.
Another option that would be even better in my opinion would be to be able to access our error log in the replay.
I strongly agree with being able to see debug data (for our bot only of course) for all matches.
I also strongly agree with that. To add an other reason, I ( and surely other players) make my algos print some insight of their decision process and it would be a great help to be able to read them for ranked matches.
For debug purposes, @kkroep, I use something I described in this topic. I have found this easier than using the playground for the very reason you mentioned.
Adding the log to replay files and including a debug viewer is something that is on our radar, but would take some effort to implement, and since replay files are public may expose information that users preferred not to expose. We are looking into solutions for this as it is an issue that has come up before, but have a few higher priorities for now.
Allowing crashed-out algos to play in playground seems like it should be a low effort change and I agree that it would be helpful so ill see if we can throw it into this week’s schedule. I could be underestimating the effort required though.
Maybe it would be easier to email the traceback to the owner’s email?