code berg activity tracker only tracks PUSHES, not commits. #2201

schatt commented 2025年11月07日 19:30:48 +01:00

Comment

The activity tracker on my profile only shows pushes - not commits. So, if I am on the main branch of a repo, do 10 commits, and then push to codeberg, it shows only one "activity" for the day.

Not a major issue, but not exactly how I'd handle activity in git.

### Comment The activity tracker on my profile only shows pushes - not commits. So, if I am on the main branch of a repo, do 10 commits, and then push to codeberg, it shows only one "activity" for the day. Not a major issue, but not exactly how I'd handle activity in git.

fnetX commented 2025年11月09日 01:37:16 +01:00

The challenge with considering commits would be that they can be falsely attributed (e.g. another account could use your email address. If the commits are not signed, it's hard to tell them apart). Also, users have vastly different workflows. Do you do one commit for a large chunk of work, or many small ones? What if you only push commits from others? Push a wrong branch that contained many commits?

I believe that pushes are a more meaningful activity, because it is the active event that you share something. No matter how large or small the chunks of work are, I believe the procedure of sharing something is an activity and probably the best one to count.

I understand that this heuristic does not always provide excellent results, but I fear that all other ways to estimate the work are either not giving better results, or are very complex to implement. And if it's too complex to understand, I don't think users will be happy, but rather confused because "I pushed two things and only one was counted" (and you'd need to explain that the heuristic didn't map the one thing to your account or it was below some threshold and just considered a fixup etc etc).

The challenge with considering commits would be that they can be falsely attributed (e.g. another account could use your email address. If the commits are not signed, it's hard to tell them apart). Also, users have vastly different workflows. Do you do one commit for a large chunk of work, or many small ones? What if you only push commits from others? Push a wrong branch that contained many commits? I believe that pushes are a more meaningful activity, because it is the active event that you share something. No matter how large or small the chunks of work are, I believe the procedure of sharing something is an activity and probably the best one to count. I understand that this heuristic does not always provide excellent results, but I fear that all other ways to estimate the work are either not giving better results, or are very complex to implement. And if it's too complex to understand, I don't think users will be happy, but rather confused because "I pushed two things and only one was counted" (and you'd need to explain that the heuristic didn't map the one thing to your account or it was below some threshold and just considered a fixup etc etc).

schatt commented 2025年11月09日 19:29:28 +01:00

Perhaps the pushes that are tracked should be on all branches then? I think right now, the help information says that it only tracks commits on the main branch.

Perhaps the pushes that are tracked should be on all branches then? I think right now, the help information says that it only tracks commits on the main branch.

arnauNau commented 2025年12月23日 16:41:13 +01:00

I agree with @schatt , usually I only push when I have a stable working version of my projects, but I have commits for relevant self-contained work I do and the push usually contains commits spanning multiple days. I don't like my activity graph only showing 'i worked' one day one time, when it was multiple days, multiple times.
Having said this, I see this not being a priority (as I see it as a polish thing more than an important feature) but coming from GitHub I miss how the activity graph worked there.

Regarding all the points @fnetX makes, I see that you mean but I don't see how choosing one 'side' over the other fixes the 'problems'. It just moves the issue of maybe counting extra 'work' for some people on the graph (over-representing) to definitely counting less 'work' to most (under-representing).

I agree with @schatt , usually I only push when I have a stable working version of my projects, but I have commits for relevant self-contained work I do and the push usually contains commits spanning multiple days. I don't like my activity graph only showing 'i worked' one day one time, when it was multiple days, multiple times. Having said this, I see this not being a priority (as I see it as a polish thing more than an important feature) but coming from GitHub I miss how the activity graph worked there. Regarding all the points @fnetX makes, I see that you mean but I don't see how choosing one 'side' over the other fixes the 'problems'. It just moves the issue of maybe counting extra 'work' for some people on the graph (over-representing) to definitely counting less 'work' to most (under-representing).

fnetX commented 2025年12月23日 20:56:07 +01:00

I never said that one side is better. But doing implementation effort just to move to another side without fixing the problems is not something we can do either.

I never said that one side is better. But doing implementation effort just to move to another side without fixing the problems is not something we can do either.